UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Influence of particle size distribution on fluidized bed hydrodynamics Ip, Trevor Tsz-Leung 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A7 I62.pdf [ 7.59MB ]
Metadata
JSON: 831-1.0058825.json
JSON-LD: 831-1.0058825-ld.json
RDF/XML (Pretty): 831-1.0058825-rdf.xml
RDF/JSON: 831-1.0058825-rdf.json
Turtle: 831-1.0058825-turtle.txt
N-Triples: 831-1.0058825-rdf-ntriples.txt
Original Record: 831-1.0058825-source.json
Full Text
831-1.0058825-fulltext.txt
Citation
831-1.0058825.ris

Full Text

I N F L U E N C E O F P A R T I C L E SIZE D I S T R I B U T I O N O N F L U I D I Z E D B E D H Y D R O D Y N A M I C S Trevor Tsz-Leung lp B. E . Sc. (Chemical Engineering) University of Western Ontario A THESIS SUBMITTED IN PARTIAL FULF I LLMENT O F T H E REQUIREMENTS FOR T H E D E G R E E OF M A S T E R OF APPL I ED SCIENCE ( C H E M I C A L ENGINEERING) in T H E FACULTY OF G R A D U A T E STUDIES CHEMICAL ENGINEERING WTe accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH COLUMBIA November 1988 © Trevor Tsz-Leung lp. 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department Date KW 13 ; l^gg DE-6 (2/88) Abstract Past literature has shown that the production efficiency of a fluidized bed can be affected by changing the particle size distribution. The hydrodynamics of fine particle fluidization were studied with FCC and glass bead powders which have different surface-volume mean particle diameter (40-110 fira.) and particle size distributions (narrow cut, wide cut and bimodal) under ambient conditions. Increasing the mean particle size in-creases the minimum fluidization velocity, minimum bubbling velocity and dense phase velocity (Ud) while decreasing the voidages at minimum fluidization and minimum bub-bling and the dense phase voidage (ej) as well as the fractional bubble free bed expansion. Increasing the particle size spread increases Ud and decreases for F C C , but no clear conclusion can be made for glass bead powders. Increasing the static bed height de-creases Ud and €d of F C C powders though it has no effect on minimum fluidization and bubbling properties. The magnitude of pressure fluctuations increases with increasing superficial gas velocity and as the size spread of the F C C powder becomes more narrow. However, the frequency of fluctuations is independent of each of these factors. Therefore, the quality and production efficiency of the fluidization process should improve with the use of a wide and continuous size distribution powder. u Table of Contents Abstract ii List of Tables vii List of Figures x Acknowledgement xiv Dedication xv 1 Background 1 1.1 Classification of Powders 1 1.1.1 Group A Powders 3 1.1.2 Group B Powders 3 1.1.3 Group D Powders ' 3 1.1.4 Group C Powders 3 1.2 Regimes of Fluidization . . 4 1.2.1 Minimum Fluidization . 4 1.2.2 Bubble-Free Bed Expansion . . . 8 1.2.3 Bubbling Regime 9 1.2.4 Slugging and Pressure Fluctuations 13 1.2.5 High Velocity Fluidization Regimes 16 1.3 Dense Phase Properties 16 1.3.1 Collapse Test 16 iii 1.3.2 Correlations for Dense Phase Properties 19 1.3.3 Effect of Particle Size Distribution on Dense phase Properties . . 21 1.3.4 Effect of Static Bed Height on Dense Phase Properties 23 1.4 Objectives of this Project 24 2 Apparatus 25 2.1 Flow Regulators 25 2.2 Solenoid Valves 28 2.3 Plenum Chamber (Windbox) 28 2.4 Distributor Plate 28 2.5 Pressure Measurement Plate 29 2.6 Main Glass Column 29 2.7 Expansion Zone 30 2.8 Freeboard 3 0 2.9 Air Filter 31 2.10 Humidity Measurement , • 31 2.11 Powder . • * 31 3 Experimental Method 32 3.1 Powder Preparation 32 3.2 Particle Density Measurement 34 3.3 Particle Diameter 38 3.4 Low Velocity Fluidization 45 3.5 CoUapse Test 46 3.6 Pressure Fluctuations 47 3.7 Sources of Experimental Errors 48 iv 4 Results and Discussion: Low Velocity Fluidization 49 4.1 Minimum Fluidization Velocity 49 4.1.1 Fluid Cracking Catalyst 49 4.1.2 Glass Beads 55 4.2 Bubble-Free Bed Expansion 59 4.2.1 Fluid Cracking Catalyst 59 4.2.2 Glass Beads . 62 4.3 Minimum Bubbling Properties 65 4.3.1 Fluid Cracking Catalyst 65 4.3.2 Glass Beads 70 4.4 Summary 74 5 Results and Discussion: Dense Phase Properties 75 5.1 Effect of Mean Particle Diameter of Fluid Cracking Catalyst 77 5.2 Effect of Particle Size Distribution for Fluid Cracking Catalyst 83 5.3 Effect of Particle Size Distribution for Glass Beads 89 5.4 Examination of Correlations for Dense Phase Properties 94 5.5 Effect of Static Bed Height on Dense Phase Properties 98 5.6 Summary 103 6 Results and Discussion: Pressure Fluctuations 104 6.1 Mean Pressure in the Fluidized Bed 104 6.2 Magnitude of Pressure Fluctuations 107 6.3 Frequency of Pressure Fluctuation I l l 6.4 Summary 112 7 Conclusions and Recommendations 115 v Nomenclature 118 References 120 Appendices 128 A M i n i m u m Fluidization Plots 128 B Terminal Velocities 146 C Raw Data from Collapse Tests 150 vi List of Tables 3.1 Skeletal Density of Fluid Cracking Catalyst as Measured by Liquid Dis-placement With Different Liquids. . 35 3.2 Particle Densities for Fluid Cracking Catalyst (FCC) and Glass Beads. . 37 3.3 Composition and Surface-Volume Mean Diameter of Different Size Distri-butions of F C C 39 3.4 Composition and Surface-Volume Mean Diameter of Different Size Distri-butions of Glass Beads 40 4.1 Minimum Fluidization Data for Fluid Cracking Catalyst. Static bed heights are 3.3 and 4.4 times the bed diameter 51 4.2 Experimental Values of Minimum Fluidization Voidage for F C C at Differ-ent Static Bed Heights 52 4.3 Predicted Minimum Fluidization Velocities for FCC Powdefs 54 4.4 Experimental Minimum Fluidization Data for Glass Beads 56 4.5 Predicted Minimum Fluidization Velocities for Glass Bead Powders. . . . 58 4.6 Bed Voidages at Minimum Bubbling and Corresponding Fractional Bubble-Free Bed Expansion Data for Different F C C Size Distributions at Different Static Bed Heights 60 4.7 Bed Voidage at Minimum Bubbling and Corresponding Fractional Bubble-Free Bed Expansion for Glass Bead Powders 63 4.8 Experimental Minimum Bubbling Properties for Different F C C Size Dis-tributions 66 vu 4.9 Ratios of Minimum Bubbling Velocities to Minimum Fluidization Veloci-ties for Different F C C Size Distributions 68 4.10 Predicted Values of Minimum Bubbling Velocities and Voidages for Dif-ferent Size Distributions of F C C 69 4.11 Experimental Minimum Bubbling Data Glass Beads 71 4.12 Predicted Minimum Bubbling Velocities and Voidages for Different Size Distributions of Glass Beads 73 5.1 Dense Phase Properties in Vigorously Bubbling Beds for Different Size Distributions of F C C 82 5.2 Experimental and Predicted Dense Phase Voidages in a Vigorously Bub-bling Bed for Different Size Distributions of Glass Beads 92 5.3 Experimental and Predicted Superficial Dense Phase Gas Velocities in a Vigorously Bubbling Bed for Different Size Distributions of Glass Beads. 93 5.4 Predicted Dense Phase Voidages in a Vigorously Bubbling Bed for Differ-ent Size Distributions of F C C 95 5.5 Predicted Superficial Dense Phase Gas Velocities in a Vigorously Bubbling Bed for Different Size Distributions of F C C . . 96 5.6 Predicted Values of Index n, a Dense Phase Parameter, in a Vigorously Bubbling Bed for Different Size Distributions of F C C and Glass Beads. . 97 5.7 Indices for Variation of Dense Phase Voidage (f3) and Superficial Dense Phase Velocity (7) on Static Bed Height at Different Gas Flowrates. . . . 101 6.1 Mean Pressure from Fluidization of Different Size Distributions of FCC. Pressure Tap is 27 mm above the Distributor 105 6.2 Mean Pressure from Fluidization of Different Size Distributions of F C C . Pressure Tap is 180 mm above the Distributor 106 viii 6.3 Frequency of Pressure Fluctuation for Fluidization of Different Size Dis-tributions of F C C . Pressure Tap is 27 mm above the Distributor 113 6.4 Frequency of Pressure Fluctuation for Fluidization of Different Size Dis-tributions of F C C . Pressure Tap is 180 mm above the Distributor 114 7.1 Summary on the Effect of Particle Size Distribution of the Fluidization Properties of F C C and Glass Beads 116 B. l Terminal Settling Velocity Correlations for Spheres 147 B.2 Terminal Velocities, Reynolds Numbers and indices 'n' for Different Dis-tributions of F C C . . 148 B. 3 Terminal Velocities, Reynolds Numbers and indices 'n' for Different Dis-tributions of Glass Beads 149 C. l Raw Data from Collapse Tests on Original F C C Distribution 151 C.2 Raw Data from Collapse Tests on Coarse F C C Distribution. 152 C.3 Raw Data from Collapse Tests on Intermediate F C C Distribution 153 C.4 Raw Data from Collapse Tests on Fine F C C Distribution. .., 154 C.5 Raw Data from Collapse Tests on Wide F C C Distribution 155 C.6 Raw Data from Collapse Tests on Bimodal F C C Distribution 156 C.7 Raw Data from Collapse Tests on Intermediate Glass Bead Distribution. 157 C.8 Raw Data from Collapse Tests on Wide Glass Bead Distribution. . . . . 158 C.9 Raw Data from Collapse Tests on Bimodal Glass Bead Distribution. . . . 159 List of Figures 1.1 Geldart's Classification of Powders 2 1.2 Regimes of Fluidization 5 1.3 Plot of Bed Pressure Drop versus Gas Velocity. 6 1.4 Bed Porosity versus Superficial Gas Velocity for Typical Group A Powders 11 1.5 Typical Bed Height - Time Plot for a Collapse Test. 18 2.1 Schematic Layout of the Apparatus 26 2.2 Detailed Drawing of the Fluidization Column 27 3.1 Dimensions for the Powder Collectors of the Air Classifier 33 3.2 Particle Size Distributions for F C C Powders with Narrow Size Distributions. 41 3.3 Particle Size Distributions for F C C powders with Similar Mean Surface-Volume Mean Particle Diameters 42 3.4 Particle Size Distributions for Glass Bead Powders with Narrow Particle Size Distributions 43 3.5 Particle Size Distributions for Glass Bead Powder with Similar Mean Surface-volume Mean Particle Diameters 44 5.1 Sample Plot of Bed Height Versus Time for Intermediate F C C Powder with Initial Superficial Velocity of 0.0086 m/s 76 5.2 Dense Phase Voidages for F C C Powders with Different Surf ace-Volume Mean Particle Diameters. Static bed height is 3.3 times bed diameter. . . 78 x 5.3 Bed Collapse Velocities for FCC Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 3.3 times bed diameter. . . 79 5.4 Dense Phase Voidages for F C C Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 4.4 times bed diameter. . . 80 5.5 Bed Collapse Velocities for FCC Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 4.4 times bed diameter. . . 81 5.6 Dense Phase Voidages for F C C Powders with Similar Mean Particle Di-ameters but Different Size Distributions. Static bed height is 3.3 times bed diameter 84 5.7 Bed Collapse Velocities for FCC Powders with Similar Mean Particle Di-ameters but Different Size Distributions. Static bed height is 3.3 times bed diameter 85 5.8 Dense Phase Voidages for F C C Powders with Similar Mean Particle Di-ameters but Different Size Distributions. Static bed height is 4.4 times bed diameter 86 5.9 Bed Collapse Velocities for FCC Powders with Similar Mean Particle Di-ameters but Different Size Distributions. Static bed height is 4.4 times bed diameter 87 5.10. Dense Phase Voidages for Glass Bead Powders with Similar Mean Particle Diameter but Different Size Distributions. Static bed height is 4-4 times bed diameter 90 5.11 Bed Collapse Velocities for Glass Beads Powders with Similar Mean Par-ticle Diameter but Different Size Distributions. Static bed height is 4.4 times bed diameter 91 5.12 Effect of Static Bed Heights on Dense Phase Voidages of Original F C C Distribution . . . 99 xi 5.13 Effect of Static Bed Heights on Bed Collapse Velocities of Original FCC Distribution 100 6.1 Pressure Fluctuations when Fluidizing Different F C C Powders Measured 27 mm above the Distributor 108 6.2 Pressure Fluctuations when Fluidizing Different F C C Powders Measured 180 mm above the Distributor 109 6.3 Typical Fluidized Bed Pressure Waveforms for F C C Distributions at Dif-ferent Bed Levels 110 A . l Minimum Fluidization Plots for Original Distribution of FCC. Static bed height is 3.3 times bed diameter 129 A.2 Minimum Fluidization Plots for Original Distribution of F C C . Static bed height is 4.4 times bed diameter . 130 A.3 Minimum Fluidization Plots for Coarse Distribution of F C C . Static bed height is 3.3 times bed diameter. 131 A.4 Minimum Fluidization Plots for Coarse Distribution of F C C . Static bed height is 4.4 times bed diameter . 132 A.5 Minimum Fluidization Plots for Intermediate Distribution of F C C . Static bed height is 3.3 times bed diameter. 133 A.6 Minimum Fluidization Plots for Intermediate Distribution of F C C . Static bed height is 4.4 times bed diameter. . 134 A.7 Minimum Fluidization Plots for Fine Distribution of F C C . Static bed height is 3.3 times bed diameter 135 A.8 Minimum Fluidization Plots for Fine Distribution of F C C . Static bed height is 4.4 times bed diameter 136 xii A.9 Minimum Fluidization Plots for Wide Distribution of F C C . Static bed height is 3.3 times bed diameter. . 137 A.10 Minimum Fluidization Plots for Wide Distribution of F C C . Static bed height is 4.4 times bed diameter 138 A.11 Minimum Fluidization Plots for Bimodal Distribution of F C C . Static bed height is 3.3 times bed diameter 139 A.12 Minimum Fluidization Plots for Bimodal Distribution of F C C . Static bed height is 4.4 times bed diameter 140 A.13 Minimum Fluidization Plots for Coarse Distribution of Glass Beads. Static bed height is 3.3 times bed diameter 141 A.14 Minimum Fluidization Plots for Intermediate Distribution of Glass Beads. Static bed height is 4.4 times bed diameter 142 A.15 Minimum Fluidization Plots for Fine Distribution of Glass Beads. Static bed height is 3.2 times bed diameter. . 143 A.16 Minimum Fluidization Plots for Wide Distribution of Glass Beads. Static bed height is 4.4 times bed diameter 144 A.17 Minimum Fluidization Plots for Bimodal Distribution of Glass Beads. Static bed height is 4.4 times bed diameter 145 Xlll Acknowledgement Professor J . R. Grace, Dr. G. K. Khoe and the staff in the Department of Chemical Engineering are greatly appreciated for their technical advice and assistance in the com-pletion of this project. Special thanks are given to my wife, Julie S. H. Ip, who provides me with tremendous support and encouragement throughout the course of my studies. I also wish to thank the Natural Science and Engineering Research Council for providing the financial support for this project. Ultimate gratitude for our Heavenly Lord who makes research ever so meaningful to the existence of human kinds. xiv Brethren...forgetting what lies behind and reaching forward to what lies ahead, I press on towards the goal for the prize of the upward call of God in Jesus Christ. (Paul in Philippians 3:13-14, NASB) xv Chapter 1 Background Fluidization has been a widely employed process in the chemical and other process industries. Essentially, fluidization occurs when the weight of a bed of powder is sup-ported by the pressure drop across the bed. When this happens, the particles circulate inside the bed. The most common mode of fluidization is gas-solid fluidization. In this case, gas is passed through a bed of reacting powdered solids. When chemical reactions are carried out in fluidized beds, the rate of reaction is strongly influenced by the amount of contacting area between the gas and solids. A fluidized bed presents more reacting surface than a packed bed. The amount of contacting surface for a fixed amount of solid can be further increased by using finer particles. This is why the diameter of most indus-trial fluid bed catalysts is small, typically less than 100 fim. However, i f the powder is too fine, the interparticle forces may affect the fluidization process in a negative manner. Hence, it is very important to understand the hydrodynamics of fine particle fluidization. 1.1 Classification of Powders Not all powders have the same fluidization characteristics. Some work has been done to classify powders according to the way they fiuidize. The most commonly used classifica-tion is the one proposed by Geldart (1973)' for air fluidization under ambient conditions (Figure 1.1): 1 Chapter 1. Background Figure 1.1: Geldart's Classification of Powder, Ambient Conditions (modified after dart, 1973). Chapter 1. Background 3 1.1.1 Group A Powders The minimum fluidization velocity, C/m/, is denned as the gas velocity at which the bed just starts to fluidize. The minimum bubbling velocity, Umb, is the gas velocity at which gas bubbles first appear. The most striking characteristics of group A powders is that the bed expands homogenously when the superficial velocity is increased from Umf to Umb- Within this range of gas velocity, though the bed is fluidized and particle circulation takes place inside the bed, no bubbles appear. No phase separation occurs in the bed. This is due to some cohesivity of the powder. Because of the cohesive forces that hold the particles together, the excess gas flow passes through the interstitial space among the particles and no bubbles are formed. When Umb is reached, the attraction force is insufficient to keep bubbles from forming. 1.1.2 Group B Powders The interparticle forces for powders of this group are negligible. There is no bubble-free expansion stage. Instead, bubbles appear as soon as the bed is fluidized, i.e. Umb is the same as Umf. No circulation of particles takes place below the minimum bubbling point. 1.1.3 Group D Powders This group of powders comprises larger and denser particles, typically 1 mm in size or larger. These particles can be spouted readily. Solid mixing is relatively poor compared to Group A and Group B particles. 1.1.4 Group C Powders For very fine powders, the cohesive forces between the particles are very strong. When the minimum fluidization point is reached or sometimes even before that, channels and Chapter 1. Background 4 cracks in the bed of solids usually appear rather than the homogenous expansion which occurs in the case of Group A powder. Gas flows through these channels. Fluidization is very difficult unless the channels are broken up by some external means such as stirring or shaking. 1.2 Regimes of Fluidization Different regimes of fluidization are described here. Each regime has its particular prop-erties and appearance as shown in Figure 1.2. Empirical correlations have been developed for these properties and they will be mentioned in this section. 1.2.1 Minimum Fluidization Suppose that gas is passed vertically through a bed of particles. At low gas velocity, the particles are stationary and the bed is called a packed bed. When the gas flow is increased to the point where the pressure drop across the bed equals the weight of the powder per unit area, then the bed is said to be fluidized provided that the gas flow is uniformly distributed. A typical bed pressure drop versus superficial velocity curve is given in Figure 1.3. The minimum fluidization point is measured as the intersection of the two linear portions. The most widely used correlation for the Umf is derived from Ergun's equation which was developed from experiments on packed bed (1952) and may be written as: PMPP-PS) = 150(1-e m / )M r o y m / . | l ^ p f f i C f t , emf is the ratio of the volume occupied by the gas in the dense phase and bubble phase (excluding intraparticle voids) to the total volume at minimum fluidization. Grace (1982) eliminated Cmf using empirical findings. The simplified correlations for small and large flir flir flir flir Air (a) (b) (c) (d) (e) Packed B u b b l e - F r e e B u b b l i n g S l u g g i n g T u r b u l e n t S t a t e Expansion , F l u i d i z a t i o n F i g u r e 1.2: Regimes of F l u i d i z a t i o n Chapter 1. Background 6 APrafl Fluidized State fl - Increasing Velocity B ~ Decreasing Velocity Umf SUPERFICIAL GAS VELOCITY. Figure 1.3: Plot of Bed Pressure Drop versus Gas Velocity (modified after Grace, 1982). Chapter 1. Background 7 particles are as follows: Umf = 0.00075 * ( P p ~ P 9 ^ d ' v for Ar< 10* (1.2) V-Umf = 0.202 * ( ( P p ~ / J g ) g ^ " ) ° - 5 for Ar > 107 (1.3) Ps In essence, Umf is a function of the gas density and viscosity, the particle density and the bed voidage at minimum fluidization. Other correlations for Umf have been developed, but the independent variables for the best of these correlations are the same as those from the Ergun's equation with only the coefficients changed. A few of these are given here. Wen and Yu (1966a, 1966b) suggested the following correlation for laminar flow: U m f ~ 1650/z ( L 4 ) Baeyens and Geldart (1973) proposed the following relationship: o.ooo9(Pp - Pg)°-93V-93X-8 Umf ~ .,0.87^ 0.066 V 1 - 0 / " "a Davies and Richardson (1966) published the correlation: where dv is the number-volume mean particle diameter; dp is the mean particle size obtained from standard sieve analysis using the following relationship (Abrahamsen and Geldart, 1980a): dP = ( ^ ) ~ l (1.7) According to these correlations, if one uses the same kind of gas and particles, changing the particle size distribution will not affect Umf as long as the bed voidage and the mean particle size remain constant. However, Kunz (1970) reported that glass beads and fluid cracking catalyst particles with higher fines content (<44 ^m) tend to have a Chapter 1. Background 8 lower minimum fluidization velocity. It also becomes more difficult to measure a distinct value for Umf as the fines level goes up (Dry et al., 1983; Geldart et al, 1984). Geldart et al. (1983) also reported that the ratio of bed pressure drop to the powder weight per unit area decreased from the ideal value of unity as the fines content increased. The high fines content powder was found to behave more and more like a Group C powder. Massimilla et al. (1972) found that with the number-surface mean particle size kept constant, a FCC powder with a broad size spectrum has higher em/ and lower Umf compared to one with a narrow size spectrum. 1.2.2 Bubble-Free Bed Expansion Bubble-free bed expansion occurs only for Group A powders. This regime has been explained by Massimilla and Donsi (1976) who proposed that bed expansion takes place through the nucleation of microscopic cavities whose size is about one to ten times the particle diameter. They defined cavities as bed voids surrounded by particles bonded to each other by cohesive forces. These cohesive forces are mainly attractive capillary forces developed as a consequence of the formation of liquid bridges between contacting bodies in the presence of a vapour-containing atmosphere. A less significant component is the Van der Waals force related to the electromagnetic fluctuation phenomenon in solids. A third component consists of electrostatic forces due to electric charges on the particles (Clift, 1986). This is particularly important for particles smaller than 5 pm in diameter. Mutsers and Rietema (1977, 1984) followed up on Massimilla's explanation by sug-gesting that the powder structure during homogenous expansion obeys the relationship: where E is the elasticity coefficient of the powder structure. The elasticity coefficient depends on the porosity of the bed. When the bed is expanded by an increase of air flow, Chapter 1. Background 9 the size of an individual cavity increases. However, the number of cavities remains the same. The elasticity of the powder structure resists expansion of the bed and break-up of cavities. For increasing bed expansion, the elasticity constant remains the same as long as the maximum bed volume for that particular value of elasticity constant is not exceeded. H the limit is exceeded, some bonds that hold the microscopic cavities together will be broken and the particles will rearrange themselves to a structure with higher porosity; the value of the elasticity constant is then reduced stepwise. Mutsers and Rietema reported that the value of the elasticity coefficient depends on the particle size distribution and types of material. Addition of a small amount of fines increases the number of contact points among particles and the elasticity coefficient. Different types of particles have different shapes. Hence the elasticity coefficient of a polypropylene bed is different from that of the fluid cracking catalyst bed. Abrahamsen and Geldart (1980) derived an expression for bubble-free expansion for F C C , alumina and glass ballotini particles having mean particle diameter less than 75pm and gases with different densities and viscosities: - f L > - ' > g = 210(t/- UmJ) + (1.9) 1 - e p l - £ m / P As indicated previously, the bed porosity goes up when the superficial velocity is increased beyond Umf. But bed expansion goes through a maximum near the minimum bubbling point. At the minimum bubbling point, the attractive forces are overcome by the excess gas flow. As a result, the microscopic cavities disappear and bubbles are formed. The bed then becomes heterogenous. 1.2.3 Bubbling Regime At the minimum bubbling point, the elasticity coefficient, as proposed by Massimilla et al. (1972), is reduced to its minimum. The bed does not have any more elasticity. IS Chapter 1. Background 10 Any increase of gas flow results in bubble formation rather than further bubble-free bed expansion. With the interparticle forces disrupted, the bed contains two phases: a dense phase and a dilute phase. The dense phase contains the solid particles and the interstitial spaces while the dilute phase is mainly composed of gas bubbles, with a very small amount of solids carried into the bubbles by the jet stream issuing from the distributor plate. At gas velocities just beyond Umf, bed contraction tends to occur because the reduction of dense phase volume is more rapid than the increase in bubble holdup (Geldart et al., 1984). Eventually, e<f reaches a limiting value which is somewhere between Cmj and emb. Any further increase in gas flow leads to an increase in total bed height (due to an increase in bubble holdup), but e<f is constant or increases only slightly (Abrahamsen and Geldart, 1980b; Rowe et al., 1978; and Vries et al., 1972). In this section, only the minimum bubbling properties are discussed. Dense phase properties are treated in Section 1.3. Different methods have been used by researchers to identify Umb- Normally it should be the point at which bubbles become visible on the bed surface, but this is subjective and can be inaccurate. For example, if the gas distribution or particle size distribution is not uniform in the bed, bubbles may appear only at certain locations of the bed while it is rather quiet in other areas. Other more objective ways have been used to determine Umb- As mentioned in the previous section, the fluidized bed should be close to its maximum height at the minimum bubbling point. Some workers identify the minimum bubbling point by choosing the point where the bed height reaches a maximum (Jacob and Weimer, 1987; Geldart et al., 1984). They believe that bed contraction occurs as soon as the bed starts to bubble. On a plot of bed porosity versus superficial velocity as shown in Figure 1.4, there is a constant and linear region at gas velocities below the minimum fluidization point. Another linear region often occurs at gas velocities between the minimum fluidization and Chapter 1. Background Figure 1.4: Bed Porosity versus Superficial Gas Velocity for a Typical Group A Powder (modified after Jacob and Weimer, 1980). Chapter 1. Background 12 the minimum bubbling points. Rowe and Yacono (1976) used the intersection of these two linear regions as the minimum bubbling point. This is applicable only for group B and D powders since for group A powders the bed starts its bubble-free expansion at U m f . Correlations have been developed for the minimum bubbling point. Geldart and Abrahamsen (1978) found that for air fluidization under ambient conditions, Umb c a n be estimated using a very simple equation: Umb = 100 *dp (1.10) where Umb a n d dp  a r e m SI units. This correlation was claimed to be satisfactory as long as the fines content (<45 pm) was less than 15% by mass. If the fines content is higher than 15%, Abrahamsen and Geldart (1980a) proposed the following relationship: Umb = 2.07exp(0.716F 4 5)^g i F (1.11) /* This correlation was developed for a variety of powders, gases and operating conditions. Foscolo and Gibilaro (1984) studied the hydrodynamic interaction between a particle and the fluid in a fluidized suspension and proposed the following criterion for the onset of bubbling: [ ( ^ ) ( ^ 1 ^ ) ] o . 5 = 0 5 6 T l ( 1 _ e m ( ) ) o . s c - 1 ( 1 1 2 ) U t Pp The correlation for minimum bubbling voidage provided by Abrahamsen and Geldart (1980a) considers the particle size distribution up to only the extent of weight fraction of powder smaller than a certain diameter. The effect of particle size distribution has also been studied by others. The results are contradictory. Simone and Harrison (1980) reported that the minimum bubbling velocity increases slightly going from a powder with a narrow size spectrum to one with a broad size spectrum while the surface-volume mean particle diameter is kept constant. The method used to prepare the silica-alumina Chapter 1. Background 13 mixture was not given in their paper. De Jong et al. (1974) also reported that the minimum bubbling velocity of cracking catalyst with a broad size spectrum is slightly higher than that with a narrow size spread. The powder with a broad size spectrum was prepared by mixing the fine and the coarse fractions. It could have been a bimodal size distribution powder, depending on the size spread of the individual fractions which was not revealed. These results are contradicted by Richardson (1971) who found a slightly lower value of minimum bubbling velocity for Diakon with a broad size spread than for a narrower distribution. Massimilla and Donsi (1976) also reported that the powder with a broad size spread has higher emb and Umb than one with a narrow size spectrum. However, for reasons not given in the paper, Massimilla and Donsi did not provide the mean particle diameter of the powder. They just claimed that the mean particle diameters of the two different powders were similar, although they might have been as much as 20% apart when one considers the size range of the narrow distribution. 1.2.4 Slugging and Pressure Fluctuations As the gas flow increases, the mean bubble size also increases. Eventually, if the bed_ is deep enough and the maximum stable bubble size big enough, slugging will occur. Slugging is characterized by the periodic rise and fall of the entire bed surface and of the bed pressure drop (Grace, 1982). This pressure fluctuation is quite important in some fluidized beds because the vibration may damage structures such as baffles and tubes inside the fluidized bed. In a freely bubbling bed, the bubbles are much smaller than the bed diameter. Many individual bubbles can be found on any horizontal plane of the fluidized bed. Noorder-graaf et al. (1987) reported that the pressure fluctuations in this flow regime tend to have irregular frequency and amplitude. The magnitude of pressure fluctuations is somewhat Chapter 1. Background 14 smaller than in the slugging regime. When the bubbles become larger and have diameters comparable to that of the bed diameter, slugging occurs. In this flow regime, only a single chain of voids can be present in the bed. This results in a single dominant frequency and much more regular pressure fluctuations which also are larger in magnitude than in the bubbling regime. The pe-riod of the pressure fluctuations is usually about one to few seconds per cycle. Fan et al. (1981, 1983) have used the change in the pressure fluctuations to identify the onset of slugging and to infer the slug rise velocity. There are two theories on the cause of pressure fluctuations. Noordergraaf et al. (1987) believed that pressure fluctuations are caused by the disintegration of a rising solid slug, followed by the 'raining' of particles. When the piston-like solid slug disintegrates at the top of the fluidized bed, many particles rain downwards and have a lower contribution to the bed pressure drop. The bed pressure is at its maximum just before the slug breaks the surface. The slugs are present only beyond a certain distance above the distributor. Below this bed level is a 'freely bubbling zone' where bubbles and jets coalesce to form slugs. In this 'freely bubbling zone', the pressure fluctuations are independent of the distance above the distributor. However, in the 'slugging zone', the average amount of particles above the pressure tap decreases. Therefore the magnitude of pressure fluctuations caused by the 'piston-rain transition' of these particles would decrease with increasing distance above the distributor. Fan et al. (1981) has a different explanation for the pressure fluctuations. In the 'slugging zone' of the fluidized bed, the pressure fluctuations are caused by the motion of the bubbles around the pressure tap. The pressure reaches maxima and minima when the roof and the floor of the bubbles reach the pressure tap, respectively. In the 'freely bubbling zone', the pressure fluctuation is caused by a combination of the following: the jet flow and the formation of bubbles which transmit the pressure fluctuations upward Chapter 1. Background 15 and the formation of the large bubbles which transmits the pressure fluctuations down-ward. The last factor accounts for the majority of the pressure fluctuations, though that contribution decreases when one gets closer to the distributor. Therefore the pressure fluctuation increases as the pressure probe moves away from the distributor and reaches a maximum at the point where slugs are formed. This is different from what Noordergraaf suggested. However, both workers concluded that the pressure fluctuation decreases in magnitude when the probe is farther away from the distributor in the 'slugging zone'. Svoboda et al. (1984) reported that the nature of pressure fluctuations in a fluidized bed is a complex function of particle properties, bed geometry, bed pressure, and prop-erties, and flow conditions of the fluidizing fluid. This study concentrates on the effects of the particle size distribution and the mean particle diameter on the amplitude and the frequency of the pressure distribution as well as the mean pressure in the fluidized bed. Svoboda (1984) found that the mean pressure in the fluidized bed increases with increasing gas velocity and particle size. The magnitude of the pressure fluctuations increases with increasing superficial gas velocity and mean particle diameter (Satija and Fan, 1985); Fan et al., 1983; and Kang et al., 1967). This is caused by the increase of slug rise velocity and void volume. The dominant frequency of pressure fluctuations decreases with increasing mean particle diameter (Lirag and Littman, 1966; Sadasivan et al., 1980; Svoboda et al., 1984; Satija and Fan, 1985). However, the frequency is either only slightly dependent on or totally independent of the superficial velocit}', as long as the gas flow is high enough that slugging occurs (Verloop and Heertjes, 1974; Sadasivan et al., 1980; and Noodergraaf et al., 1987; and Satija and Fan, 1985). Morse and Ballou (1951) found that the fluidization quality improves when a bed of solids has a broad size spectrum. This implies smaller pressure fluctuations and more predominant dense phase gas flow. The appearance of the slugs is also influenced by mean particle size. Kehoe and David-son (1970) found that for fine particles (<70 ^m), slugs are symmetrical. However, wall Chapter 1. Background 16 slugs were observed for coarse particles. The bed height and particle density also affect the pressure fluctuations in a fluidized bed. However, they will not be investigated in this work. 1.2.5 High Velocity Fluidization Regimes Turbulent fluidization occurs when the gas velocity is increased beyond a certain transi-tion velocity. During turbulent fluidization, the bed surface is quite distinct and stable. Small voids darts to and fro. Pressure fluctuations are smaller in amplitude and higher in frequency than during slugging. When the gas velocity is increased further to the transport velocity and higher, the particles are simply blown out of the top of the reactor. There is no longer any upper surface of the bed. Solids must be added to the reactor to prevent it from being emptied. 1.3 Dense Phase Properties The dense phase of a fluidized bed is very important. This is where the gas and the solid establish intimate contact. Hence chemical reaction takes place primarily here. It is therefore vital to study the superficial dense phase gas velocities and dense phase voidages. These dense phase properties can be measured using the collapse test technique developed originally by Rietema (1967). The collapse test is of particular importance for Group A powders because of their slow de-aeration rate. In this section, the collapse test is explained. Empirical correlations for the dense phase properties are also discussed. 1.3.1 Collapse Test Tung et al. (1989) proposed that in a collapsing gas-solid fluidized bed after the gas flow has been suddenly cut off, the gas flow can be divided into three components: l)via Chapter 1. Background 17 bubble translation and throughflow; 2)through the interstices of suspended particles; 3)driven out by the consolidating particles. During the collapse of the bed of particles, all three types of gas flow take place simultaneously. After the gas supply to a bubbling bed is shut off, bubbles ascend through the bed and leave through the surface of the bed. This is the bubble escape stage. The bed level drops quickly during this period. At the end of the bubble escape stage, the gas flow through the interstices of suspended particles becomes predominant. This is the hindered sedimentation stage. The bed collapse rate is constant and slower than for the bubble escape stage. The bed is divided into two layers. Particles are piling up at the bottom of the bed. As a result, the bottom layer is denser than the" top layer. The boundary between the two layers continues to rise until it reaches the top. That marks the end of the hindered sedimentation stage and the beginning of the solid consolidation stage. During the solid consolidation stage, the bed has uniform density. The gas in the interstitial space is expelled hy the accumulation of particles. The rate of bed collapse is relatively slow. The collapse test provides data for a plot of height versus time. An example is shown on Figure 1.5. Abrahamsen and Geldart (1980b) indicated that the middle, linear part of the curve corresponds to the hindered sedimentation stage. The slope of that part is the superficial dense phase gas velocity of the fluidized bed. If this linear part is extrapolated to time zero, where the intercept on the ordinate axis is the dense phase bed height. The dense phase voidage can then be calculated. Abrahamsen and Geldart (1980b) also reported that the rate of bed collapse depends on undesirable factors such as the plenum chamber geometry and the pressure drop across the distributor. Adjustments may have to be made to either the experimental design or experimental data in order to obtain the true values for the dense phase properties. Chapter 1. Background Consolidation Time Figure 1.5: Typical Bed Height - Time Plot for a Collapse Test (modified after Ab hamsen and Geldart, 1980b). Chapter 1. Background 19 1.3.2 Correlations for Dense Phase Properties The dense phase properties of a bubbling bed have been of great interest to researchers. Different correlations have been developed to predict Ud and ej. Some are simpler than others, and some also take into consideration the fines content. One of the oldest the-ories of fluidization is the 'two-phase theory of fluidization', proposed by Toomey and Johnston (1952), and Davidson and Harrison (1963), which assumes that Ud and td in a bubbling bed are the same as their respective values at minimum fluidization. However, Geldart (1986) indicated that numerous results have shown that the experimental values of Ud and td exceeds those predicted using the two-phase theory. An old correlation for the dense phase properties was derived from liquid-solid flu-idization data collected by Richardson and Zaki (1954). The experiments were conducted under ambient conditions using spherical glass ballotini particles and various organic so-lutions. The correlation is as follows: where n = 4.65 for Ret < 0.2 n = (4.35 + 17.5^)Re; 0 0 3 0.2 < Ret < 1 n = (4.45 + 18^)#e t - 0 1 1< Ret < 200 n = 4.45i2et-0-1 200 < Ret < 500 n = 2.39 Ret > 500 Though this correlation was developed from solid-liquid fluidization, it has been applied for gas-solid fluidization for superficial velocities between Umf and Umb with reasonable accuracy (Davies and Richardson, 1966; Abrahamsen and Geldart, 1980a). Chapter 1. Background 20 Kmiec (1982) also did some solid-liquid fluidization experiments using ion-exchange particles, agalit, glass beads and water. He found that was well represented by: (lSRep + 2 .7^- 6 8 7 ) 0 - 2 0 9 = A r o . 2 0 9 (L 1 4 ) where AT = R p-Pg(pP - Pg)9<l3 sv pi Kmiec claimed that his correlations produced results similar to those from Richardson and Zaki's equation. Foscolo and Gibilaro (1983) examined the hydrodynamics from the packed bed state to the fully expanded state and obtained the correlation: Ud _ [0.0777flet(l + 0.0194iZet)e^8 + l ] 0 5 - 1 . . Ut ~ 0.0388i?et ^ ' °^ Abrahamsen and Geldart (1980b) considered the dense phase properties for gas-solid fluidization of group A powders. The particles were ballotini, alumina and cracking catalyst. Different types of gas (argon, air, carbon dioxide, freon and helium) were used in order to study the effect of gas density and viscosity on dense phase properties. Fines fraction, F 4 5 , defined as the weight fraction of particles with diameter (from sieve analysis) less than 45 pm, was also considered. The correlation is as follows: 1 - e^f 2.54p° o lVO O 6 6exp(0.090F 4 5) l - e d d^g^{pp-Pgf^Hmf3 (1.16) ( _ ^ _ ) 0 . 7 = ( j L ) 3 ( J _ £ m / ) (1.17) Umf £mf 1 — Crf From these correlations, one can see that if the mean particle size is constant, increasing the fines fraction can make the dense phase voidage go up by a factor of exp(0.09F45). Chapter 1. Background 21 These correlations were claimed to be able to predict the dense phase properties within 20% of their experimental values. The correlations were also shown to fit well with data from the literature by other authors. Dry et al. (1983) performed similar experiments with iron oxide, carbon powder, cracking catalyst and air. Their fines were denned as those particles with volume mean diameter less than 22 pm. The correlations are as follows: U d = 22607"" ( L 1 8 ) l - e r f = 0.212exp(1.13exp(-2.07F22)) (1.19) These correlations did not agree very well with earlier data from the literature. Dry et al. had problems determining the particle density of the porous cracking catalyst, and this may have affected the reliability of their correlations. 1.3.3 Effect of Particle Size Distribution on Dense phase Properties Abrahamsen and Geldart (1980b) found that the dense phase behavior depends strongly on the weight fraction of fines defined as particles smaller than 45 pm in diameter. Their correlation suggests that the overall particle size distribution is not important in determining the dense phase properties, provided the fines fraction and the mean particle size (measured with standard sieve analysis) are constant. Their correlations also suggested" that ej increases when the fines fraction F45 increases for a constant mean particle diameter. Simone and Harriott (1980) also found that the particle size distribution does not affect the dense phase expansion in a vigorously bubbling bed when the surface-volume mean diameter of the powder is kept constant. They also found that the fractional dense phase expansion in a vigorously bubbling bed, Ud/Umj, is always about 40-50% of the •fractional bed expansion at minimum bubbling, UmbfUmf-Chapter 1. Background 22 Geldart and Wong (1985) used powders with certain degrees of cohesiveness to study the dense phase properties. Their alumina powder had a mean particle diameter of about 30 pm (obtained with sieve analysis), so the powder is still in Group A according to Geldart's classification. Geldart and Wong made corrections for the volume of air -trapped in the plenum chamber when the collapse test is started. It was found that increasing the fines fraction (Fih) by 10% would increase the superficial dense phase velocity by about 50%. The difference in the mean particles diameter between the two batches of powder used was only 2 pm which is too small to account for the big difference in the dense phase velocity. The studies reviewed above appears to be the only ones in which the effect of particle size distribution on dense phase properties has been investigated with constant mean particle size. The other work appear to involve uncontrolled studies in which both the fines fraction and the mean particle size were varied at the same time. Some of this work is discussed briefly here. As mentioned before, Dry et al. (1983) claims that the fines fraction (F22) is directly responsible in the determination of the dense phase voidage of fluidized bed filled with iron oxide/carbon mixture or fluid cracking catalyst. However, in both of their experiments, the mean particle size was not kept constant. Barreto et al. (1983) claimed that when the fines fraction (i^s) increases by 46%, the dense phase voidage of the zeolite bed goes up by about 10%. However, they made the same mistake of neglecting the decrease in mean particle diameter. It is shown in Figure 1.4 that at gas velocities beyond Umb, the overall bed voidage decreases initially but eventually increases again with increasing superficial velocity. How-ever, the dense phase voidage continues to decrease and levels off to a certain value even though the overall bed voidage keeps going up (Abrahamsen and Geldart, 1980b). Rowe and Yacono (1976) found that while this means a small drop in Uj in the early stage of Chapter 1. Background 23 the bubbling flow regime, the general trend is that the permeability of the dense phase continues to increase despite the nearly constant dense phase voidage. This is indicated by the upward trend of the superficial dense phase gas velocity with increasing gas flow, even when the superficial gas velocity is ten to fifteen times Umf. This trend is more apparent when the fines fraction (FAS) is higher than 20% with a mean particle diameter of about 45 pm. However, other workers such as Abrahamsen and Geldart (1980b) and Dry et al. (1983) suggested that the increase or decrease of Ud should follow the variation of ed, even with powders having high fines fractions. Hence, Ud approaches a constant value with increasing gas flow in a vigorously bubbling bed. 1.3.4 Effect of Static Bed Height on Dense Phase Properties There is some controversy regarding whether or not dense phase properties are functions of the static bed height. Some workers have found that the dependence of superficial dense phase velocity and dense phase voidage on the static bed height is negligible (Vries et al., 1972; Bohle and Swaay, 1978; and May, 1959). Abrahamsen and Geldart (1980b) indicated that the dense phase voidage changes along the height of the bed. The region closest to the distributor plate had the highest voidage while the top region had a dense phase voidage approaching its value at minimum fluidization. Since the superficial dense phase gas velocity is related to the dense phase voidage, the latter is also a function of bed height. The results were collected with the bed height ranging from 0.3 to 0.9 m, yielding: Ud a tf0-0-244 (1.20) Drj' et al. (1983) showed a weak dependence of dense phase voidage on bed height. Their results were based on a bed depth in excess of 2 m and led to: ' e d a F - 0 0 3 9 (1.21) Chapter 1. Background 24 1.4 Objectives of this Project The correlations described above indicate that the mean particle diameter, usually the surface-volume mean, plays a very important role in the determination of minimum fluidization, bed expansion, minimum bubbling and dense phase properties. However, the influence of particle size distribution has usually been neglected. Some workers have investigated the effect of particle size distribution on different aspects of fluidization. However, they were usually uncontrolled studies with the mean particle diameter not kept constant but changing when fines or coarse particles were added to the powder. These results cannot be used reliably to draw any valid conclusions. For example, when Geldart (1972) replotted Matheson's (1949) results with the changes in particle diameter taken into consideration, it was concluded that the Stormer viscosity of the fluidized bed does not depend on the particle size distribution, exactly opposite to the original conclusion. This shows that it is vital to consider changes in mean particle diameter when studying the effect of particle size distribution. The powders used in the present work belong to Group A, the type of powder most frequently used for catalytic reactions in the chemical industry. The objective of this work was to study the effects of surface-volume mean particle diameter and the complete particle size distribution on different aspects of fluidization. The areas covered include minimum fluidization, bubble-free fluidization, minimum bubbling, dense phase prop-erties and pressure fluctuations. The surface-volume mean particle diameter was kept constant while the effects of the particle size distribution were studied. Chapter 2 Apparatus The experimental equipment was composed of gas flow regulators, a plenum chamber (windbox), a distributor plate, pressure measuring devices, a main glass column, an expansion section, a freeboard and an air filter. A layout of the whole apparatus and a detailed drawing for the components of the fluidization column are shown in Figures 2.1 and 2.2, respectively. 2.1 Flow Regulators The main flow regulators used in this project are a P V C diaphragm valve and a needle valve attached to a. rotameter. These are high precision valves which enable the oper-ator to control the flow very accurateljr. Two rotameters are used to indicate the fluid flowrate. Rotameter RI, made by Brooks Instruments, has a tube number of R-6-25-1-A. It contains a glass float and a steel float for maximum flowrates of 2.1*10 -4 and 3.8*10 -4 m 3/s, respectively, under conditions of 20°C and 1 atm. Rotameter R2, manufactured by Porter Instrument, has tube number B-175-60 and contains a sapphire float for maximum gas flowrate of 8.5*10 -5 m 3/s under the same conditions. A pressure regulator 'PR' was used to lower the air pressure from that of the building air to a pressure that the appa-ratus can stand. The air coming through 'RI' can go into the stream that contains 'R2' or be vented. This is controlled by the three way valve 'V3'. Gate valve 'V4' provides a path by which the air can bypass the rest of the equipment. This is to avoid a pressure buildup during the collapse test. Needle valve 'V5' is used to control the air efflux from 25 Chapter 2. Apparatus 26 R2 AIR OUT V2 V4 V3) RI vt RIR IN ORY BULB (WO UET BULB TEnPERATURES filR OUT V5 SV2 FREEBOflRO EXPANSION ZONE MANOMETER DISTRIBUTOR -SV1 MAIN GLASS COLUMN SEE UIN080X PRESSURE "TRANSDUCER —MANOMETERS Figure 2.1: Schematic Layout of the Apparatus. Chapter 2. Apparatus 27 ->AIR TO FILTER BAG U 0.229 i FREEBOARD 0.305 a PRESSURE TAP. , 0.10m , EXPANSION ZONE 0.100 • MAIN COLUMN AIR IN 0. 21 ra 1.83 n 0.031 • —*-• 0.102 ra Figure 2.2: Detailed Drawing of the Fluidization Column. Chapter 2. Apparatus 28 the plenum chamber when solenoid valve 'SV2' is opened. 2.2 Solenoid Valves Solenoid valves are electrically triggered valves which are either fully open or fully closed. Solenoid valve 'SV1' is normally closed (closed when no electrical current flows through the valve) while 'SV2' is normally open. The advantage of a solenoid valve over the hand-operated valve is that it takes only miUiseconds to fully open or close a valve. Both valves are triggered simultaneously with a single switch. When the valves are turned on, air may flow' from the main air supply to the plenum chamber and up through the distributor, but air'cannot escape through 'SV2'. When the valves are off, the air supply to the plenum chamber is stopped. Air can escape from the plenum chamber through 'SV2' to leave the apparatus provided that there is a pressure differential between the two ends. 2.3 Plenum Chamber (Windbox) The purpose of the steel plenum chamber is to develop a uniform pressure below the distributor plate so that gas flow is uniform through the distributor. Air enters through a 25 mm inlet pipe and leaves through the 100 mm diameter distributor into the glass column above. Two ports are found on the sides of the bottom cylindrical section. One of them serves as a pressure tap while the other is for letting air out through solenoid valve 'SV2'. 2.4 Distributor Plate Some of the particles used in this project have diameters as small as 5 pm. A stainless s.teel perforated plate with one millimeter openings was originally used as distributor, but Chapter 2. Apparatus 29 it was unsuccessful because the powder could not be prevented from dropping into the plenum chamber while the column was being cleaned with the air supply shut off. Two pieces of chromatography cardboard were then used. This provided a larger pressure drop across the distributor than the perforated plate. The pressure drop across the distributor plate was at least 25% of the bed pressure drop for the range of gas flowrates used in the experiment. Gas distribution with the cardboard distributor was found to be very uniform. 2.5 Pressure Measurement Plate A steel pressure measurement plate 'PMP1' is located right above the distributor plate. This plate provides two ports for pressure measurement. One port is for measuring the pressure drop across the bed of powder, while the other is for measuring the pressure drop across the distributor. All pressure taps are connected by plastic tubing to water or mercury manometers depending on the magnitude of the pressure gradient. Another pressure measurement plate 'PMP2' was used to measure the pressure fluc-tuation in the fluidized bed in the bubbling and slugging flow regimes. This was placed at either 27 mm or 180 mm above the distributor plate. It provides an opening for a 6.3 mm steel tube. A brass sintered filter of pore size 7 pm is attached to the inner end of the steel tube so that no particles block the opening. The outer end of the steel tube is connected to a DISA capacitance type pressure transducer. 2.6 M a i n Glass Column A glass column made by Pyrex Glassware contained the fluidized bed. Glass is superior to other materials such as plexiglass and steel for two reasons. Firstly, glass is transparent making it much easier to detect visually what is taking place inside the column. One Chapter 2. Apparatus 30 can also use a video-camera to record the experiment. Secondly, the particles have less tendency to adhere to the inside of the glass column than to a plexiglass column. The particles used for the experiments are very fine, most being smaller than 100 pm. Static electricity can be a major problem if a plexiglass column is used. The main glass column has an inner diameter of 100 mm, a wall thickness of 13 mm and a column height of 1.83 m. The inner diameter is uniform thoughout the length of the column. A metric tape was attached to the outside of the column to indicate bed levels. 2.7 Expansion Zone A steel conical section connected the 100 mm glass column to the freeboard which is 230 mm in diameter. The slanting part has a slope of 60 degrees to the horizontal. Any particles with an angle of repose of less than that should roll down to main glass column. A pressure tap is located in this section so that one can determine the system pressure close to where the air leaves the apparatus. 2.8 Freeboard At high gas flowrates, particles can be blown out of the main column. The freeboard is used to return these particles back to the main column. The freeboard is a short cylindri-cal glass.section made of Q V F glassware. The inner diameter and the wall thickness are 0.23 and 0.013 m respectively. The height of the section is 0.30 m. Since the freeboard has a diameter 2.3 times larger than that of the main column, the air velocity drops by a factor of 5 going from the main column to the freeboard. Particles with terminal settling velocities higher than the air velocity in the freeboard tend to be returned to the main column. Chapter 2. Apparatus 31 2.9 A i r Filter A paper vacuum bag is placed on top of the freeboard to trap any particles that escape the freeboard. This ensure that virtually no particles are lost from the system. 2.10 Humidity Measurement Two thermometers placed side by side are inserted into the inlet piping. The mercury bulb on one of these thermometers is wrapped with gauze and attached to a water reservoir to ensure a constant supply of water. The wet bulb temperature is obtained from this thermometer. The other thermometer indicates the dry bulb temperature of the air entering the fluidized bed. 2.11 Powder Two kinds of powder were used for our experiments. Thejr are spent fluid cracking cata-lyst (FCC) obtained from the ESSO refinery at loco, B.C. and glass beads manufactured by the Potter company in New Jersey. The properties of these powders are discussed in Chapter 3. Chapter 3 Experimental Method The experimental procedures include preparation of the powders, measurement of the physical properties of the powders, low-velocity fluidization, collapse tests and pressure fluctuation measurements. 3.1 Powder Preparation The original spent fluid cracking catalyst (FCC) was separated into six different size fractions using an air classifier in the Mining and Mineral Process Engineering Depart-ment at University of British Columbia. The air classifier consisted of six steel conical chambers arranged in series, and the dimensions of these are given in Figure 3.1. The separation involved two passes. The respective air flowrates for the first and second passes were 0.0021 and 0.0022 m/s 3, respectively. The original bulk F C C powder was charged into the air classifier during the first pass. The fraction collected in cone number 3 was then passed into the classifier during the second run. Four size tractions of F C C labelled Original', 'coarse', 'intermediate' and 'fine' were collected for our experiments. 'Original' distribution was the unseparated F C C powder. 'Coarse' and 'fine' fractions were the powder collected in cones number 2 and 4, respectively, from the first pass. 'Intermediate' fraction was collected in cone number 3 from the second pass. No separation was necessary for the glass beads because they came in three discrete size fractions. 32 Chapter 3. Experimental Method flir + Fine Portion of Solid Feed i • fiir + Solid Feed A i r C l a s s i f i e r Cone Top Cross-sectional Area 1 0.00385 2 0.00709 3 0.0145 4 0.0284 5 0.0564 6 0.108 Figure 3.1: Dimensions for the Powder Collectors of the Air Classifier. Chapter 3. Experimental Method 34 3.2 Particle Density Measurement The skeletal density for a particle is defined as the mass of the particle per unit volume of the solid part of the particle (excluding any voids inside the particle). The particle density is denned as the mass of the particle divided by the sum of the volume of the solid and volume of the internal voids, i.e. envelope volume. It is very difficult to get an accurate measurement of the particle density. Er-gun's (1951) gas flow technique with a modification described by Abrahamsen and Gel-dart (1980a) was tried. This relies on making a packed bed of two different heights with the same batch of powder. However, the fluid cracking catalyst is not very compressible. The static bed height could only be changed by 2.5% which was not large enough to produce consistent results. Skeletal density, pt, was measured with the liquid displacement method using a spe-cific gravity bottle. A known volume of liquid is added to a certain weight of solid. The volume of solid can be obtained from the volume of liquid that the solid displaces. Then the skeletal density can be calculated. The mixture of liquid and solid had to be well stirred for 15 to 30 minutes to drive out any air bubbles trapped in the liquid. Only then can measurements be taken. In the case with water, the liquid/solid mixtures were kept for a few days with periodic shaking and the results were compared to see if there was significant volume of non-visible bubbles. The results are given in Table 3.1. The skeletal density of FCC as measured with different liquids does not show any significant differ-ence. Water is supposed to penetrate the internal pores of the particles, while carbon tetrachloride is not supposed to penetrate any. But the results show only a 3% difference in skeletal density as measured with various types of fluid. The color of the catalyst is dark grey so that it is obviously spent catalyst. The internal pores are probably coated with coke so the pores of these spent catalyst particles are likely shallower than those Chapter 3. Experimental Method 35 Table 3.1: Skeletal Density of Fluid Cracking Catalyst as Measured by Liquid Displace-ment With Different Liquids. "Solid-liquid mixture degassed for few days. Liquid Skeletal Density (kg/m3) Water 2370 Ethyl Ether 2370 Methanol 2410 Carbon Tetrachloride 2310 Water* 2330 Chapter 3. Experimental Method 36 of the fresh catalyst. Carbon tetrachloride may therefore have penetrated the pores as deeply as water. Hence the two types of liquid produce similar results. The solid-liquid mixtures showed insignificant amounts of trapped air bubbles after thirty minutes of shaking. Leaving the mixture for a few days did not make any significant difference in the measured skeletal density. The results obtained with water displacement have been chosen as the true skeletal density of thepowder. The particle density, pp, is measured with the 'wet cake' method proposed by Abra-hamsen and Geldart (1980a). Water is added to a sample of powder until the particles stick together like a cake, i.e. they are no longer free-flowing. If x is the volume of water needed to just cake one kilogram of powder, then the particle densit)' can be calculated as follows: ft = r ^ i (3-1) Fluid cracking catalyst particles are porous. So both the skeletal and particle densities have to be measured. Different sizes of particles may have different fractional internal voids. Hence it is best to measure the particle density for each fraction of the F C C . However, glass beads are non-porous so that the skeletal and particle densities are the same. Only the liquid displacement method was therefore necessary to derive the particle density for the glass beads. The results on particle density of F C C and glass beads are shown in Table 3.2. The particle density of F C C increases by 12.4% going from the coarse to the fine fractions. This is perhaps due to the larger probability of bigger particle having closed or larger pores resulting in greater internal porosity for larger particles. Chapter 3. Experimental Method 37 Table 3.2: Particle Densities for Fluid Cracking Catalyst and Glass Beads. Powder Size Distributions Particle Density (kg/m3) F C C Original 1444 F C C Coarse 1384 F C C Intermediate 1455 F C C Fine 1556 F C C Wide 1423 F C C Bimodal 1440 Glass Beads All Distributions 2450 Chapter 3. Experimental Method 38 3.3 Particle Diameter The permeametry method was used to derive the specific surface area of the particles. The details of this procedure are given in the instruction manual for the 'Quanta-Sorb' surface area analyzer. Essentially, one needs a plot of bed pressure drop versus air velocity from a packed bed of particles whose specific surface is measured. The bed is packed to about 0.3 m in height. The Carmen-Kozeny equation is then used to calculate the specific surface which in this case refers to the 'envelope' surface of the particle. The specific volume (envelope) of the particles can easily be calculated from the particle density. Dividing the specific volume by the specific surface produces the surface-volume mean diameter of the particles (Stockham, 1978). Three F C C powders with different particle size distributions but similar surface-volume mean diameter were used for the experiments. The first fraction is the 'intermedi-ate size distribution' composed solely of the intermediate fraction mentioned previously. The mean particle diameter for this intermediate fraction is then matched in two other made-up mixtures. A 'bimodal size distribution' was prepared by mixing the fine frac-tion and the coarse fraction. A third 'wide size distribution' was made up of the original catalyst plus some of either the coarse or the fine fraction so that the final surface-volume mean diameter matches that of the intermediate fraction. In the case of glass beads, the wide size distribution is a mixture of the coarse, intermediate and fine fractions. The particle diameter for different distributions of F C C and glass beads are given in Tables 3.3 and 3.4. The particle size distributions are shown in Figures 3.2-3.5. According to Geldart's powder classification, all the size distributions of F C C belong to Group A. For the glass beads, all the size distributions also belong to Group A except for the coarse fraction which falls into Group B. The bimodal F C C distribution is made up of fractions that are both Group A while the bimodal glass beads powder consists of a Group A Chapter 3. Experimental Method 39 Table 3.3: Composition and Surface-Volume Mean Diameter of Different Size Distribu-tions of F C C Size Distribution Composition d,t> Original - 45.0 Coarse - 83.3 Intermediate - 53.2 Fine - 34.1 Wide 65.6% Original +34.4% Coarse 53.1 Bimodal 67.5% Coarse +32.5% Fine 54.0 Chapter 3. Experimental Method 40 Table 3.4: Composition and Surface-Volume Mean Diameter of Different Size Distribu-tions of Glass Beads Size Distribution Composition dsv [pm) Coarse - 113.0 Intermediate - 72.7 Fine - 37.4 Wide 10.0% Fine +50.0% Middle +38.0% Coarse 71.9 Bimodal 76.1% Coarse +23.9% Fine 73.1 Chapter 3. Experimental Method 41 Figure 3.2: Particle Size Distributions for F C C Powders with Narrow Size Distributions. Chapter 3. Experimental Method 42 o 180.0 S U R F A C E V O L U M E M E A N D I A M E T E R ( M I C R O N S ) Figure 3.3: Particle Size Distributions for F C C powders with Similar Mean S face-Volume Mean Particle Diameters. Chapter 3. Experimental Method 43 20.0 40.0 60.0 80.0 100.0 120.0 110.0 160.0 180.0 200.0 220.0 S U R F A C E V O L U M E M E A N D I A M E T E R ( M I C R O N S ) Figure 3.4: Particle Size Distributions for Glass Bead Powders with Narrow Purticle Size Distributions. Chapter 3. Experimental Method 44 o SURFACE VOLUME MEAN DIAMETER (MICRONS). Figure- 3.5: Particle Size Distributions for Glass Bead Powder with Similar Mean Sur-face-volume Mean Particle Diameters. Chapter 3. Experimental Method 45 powder and Group B powder. Khoe (1988b) found that the particles from all the F C C distributions are quite free-flowing. However, the glass beads appeared to be sticky, especially the intermediate fraction and, even more so, the fine fraction. Khoe examined the powders using mi-croscopy and found that the glass bead particles from the intermediate and fine fractions stuck to each other with a web-like formation. Streaks of particles spread out in all direc-tions. The F C C particles did not show any significant cohesivity when examined under the microscope. Strong interparticle forces are characteristic of Group C powders. This means that the intermediate and fine fractions of glass beads may have some Group C characteristics although they are Group A powders according to Geldart's classification. 3.4 Low Velocity Fluidization The minimum fluidization and minimum bubbling velocities were measured for each powder. The low flow stream, measured and controlled by 'R2' and 'V2', was used to control the flow of air into the main column. The bed was filled with powder to a certain height which should be constant throughout the experiment. Solenoid valves 'SVT and 'SV2' were triggered so that they were open and closed respectively. Three-way valve 'V3' was adjusted so that air passed through the low flowrate path. Before any readings were taken, the bed of powder was well mixed for a few minutes with the air flowrate well beyond that corresponding to the minimum bubbling point. The air supply was then shut off. The bed settled to a static bed height. The air flow was then increased in small increments. For each flowrate, about two minutes were allowed to ensure a steady air flow through the column before readings were taken. The following readings were taken when steady state was reached: rotameter 'R2' reading, gauge pressure inside the plenum chamber, pressure drop across the bed and bed height. The rotameters were calibrated Chapter 3. Experimental Method 46 under atmospheric pressure. The actual volumetric gas flowrate was calculated using the plenum chamber pressure and bed pressure. The dry bulb and wet bulb temperatures of air inside the system were measured at a location between 'R2' and 'SV1' and the relative humidity was close to 50% throughout the experiments. The ambient pressure was also recorded. The point where bubbles first appear was noted. The first part of these experiments was finished when the air velocity reached two to three times the minimum bubbling velocity. The second part of the .experiments was basically the same as the first one, except that it was performed by decreasing the air velocity. Between data points, the bed of powder was well mixed at a superficial velocity about twice the minimum bubbling velocity. The same set of measurements was made for each air flowrate. The point where bubbles were last seen was noted. The results were used to generate plots of bed pressure drop and bed height versus air velocity for both increasing and decreasing air velocity. The minimum fluidization point was obtained from the plot of bed pressure drop versus decreasing air velocity. The plot should contain two linear sections as shown in Figure 1.3, the intersection of these linear sections giving the minimum fluidization point. The minimum bubbling point is taken as the average of the points where bubbles first appeared (increasing velocity) and where bubbles were last seen (decreasing velocity), these points were in general within 6% of each other. 3.5 Collapse Test Collapse tests were performed with a wide range of superficial gas velocity: from about 1.5 mm/s to 40 mm/s. As in the previous experiment, the powder was well mixed before each run. The Chapter 3. Experimental Method 47 gas flowrate was then adjusted to the desired value. For air velocities below 10 mm/s, rotameter 'R2' served as the flow indicator. For higher flowrates, 'R2' was bypassed and 'RI' became the flowrate indicator. The gauge pressure inside the plenum chamber was recorded. At the start of the collapse test, the operator turned off both solenoid valves. Hence no more air could go through 'SV1' and into the main column while 'SV2' was open. The air flow from the plenum chamber through 'SV2' to the outside was controlled by adjusting valve 'V5'. The operator adjusted 'V5' in order to minimize the pressure difference across the distributor. This pressure drop was indicated on a water manometer. The purpose of this was to limit the airflow through the distributor during the collapse test. This procedure was repeated three or four times for each gas flowrate. -The course of the bed collapse was recorded using a Sony Betamax videocamera. A stopwatch and a metric tape were placed on the side of the main column to indicate the time and the bed height. 3.6 Pressure Fluctuations The instantaneous pressure was measured at a height 27 mm above the distributor for a total bed height of about 0.5 m. The reference point for the pressure measurements was in the expansion zone. The pressure was measured using the DISA pressure transducer. The results were recorded on a chart recorder at a paper speed of 6.7 or 13.3 mm/s. Pressure fluctuations were measured for the following F C C size distributions - original, wide, intermediate and bimodal. Four different superficial velocities ranging from 0.036 to 0.259 m/s were used for each distribution of powder. Similar measurements were made at a location 0.18 m above the distributor. For these measurements, a separate glass section of height 0.15 m and diameter 0.10 m was added to the bottom of the main column. The results were also recorded on videotape Chapter 3. Experimental Method 48 for future reference. Pressure measurements were not made at other bed levels because of the difficulty in drilling holes in the glass column. 3.7 Sources of Experimental Errors The daily variations of air temperature and relative humidity were less than 2°C and 5%, respectively, throughout the experiments. The temperature of air leaving the fluidization column was within 1°C of that entering the column. No significant change (less than 5%) in the relative humidity of air was noticed when it was passed through the fluidized bed. Ideal collapse tests require an instantaneous stoppage of fluid flow into and out of the grid the moment the test starts. However, in these experiments, a small amount of gas flow through the distributor occurred for a short period of time after the bed collapse had started. It usually took the operator about 1 s to equalize the pressure across the distributor. The 'bubble escape stage' usually lasted for 2 to 4 s and the usual collapse times were about 5 to 10 s for the glass beads and 10 to 20 s for the F C C powders. Hence the error introduced was probably small. Additional errors might be introduced by the DISA pressure tranducer and the chart recorder when the fluidized bed pressures were measured. The instruments were cal-ibrated regularly before the start of each experiment. The calibration curves showed changes of less than 4%. The linearity of the instruments was excellent with coefficient of correlation of at least 0.999 in the calibration curves. Chapter 4 Results and Discussion: Low Velocity Fluidization 4.1 Min imum Fluidization Velocity The experimental results for minimum fluidization velocity are shown in this section. Minimum fluidization plots for individual size distributions are given in Appendix A. Results for different static bed heights are compared. Empirical correlations are examined to test their validity. Appendix B shows the predicted terminal settling velocities which are needed to predict some fluidization properties of the powders. 4.1.1 Fluid Cracking Catalyst The minimum fluidization points for all the F C C distributions were clear and repro-ducible. Though the surface-volume mean particle diameters for some powders examined were smaller than those used by most other workers, problems in defining the minimum fluidization conditions were never encountered in our experiment. Instead, there was always a clear transition on the plots of bed pressure drop versus gas velocity with two distinctly linear regions when the gas flow was increased or decreased. The experiments with the original and intermediate powders were repeated a few times over a period of three months. For both distributions, the variations of minimum fluidization velocity and voidage were less than ± 1 . 5 % and ±0.6% of their respective means. These variations might have been introduced by simple experimental errors such as minor fluctuations of fluid flow and errors in the bed height measurement due to 49 Chapter 4. Results and Discussion: Low Velocity Fluidization 50 uneven bed surface. Increasing the mean particle diameter increased the minimum fluidization velocity of the powder as expected. The minimum fluidization voidage was found to increase as the mean particle diameter decreased as shown in Tables 4.1 and 4.2. Among the three size distributions having essentially the same mean particle diameter (labelled 'intermediate', 'wide' and 'bimodal'), the surface-volume mean diameter and the particle density differ by as much as 1.7% (0.9pm) and 2.2%, respectively, which can theoretically result in about 5% changes in Umf (estimated using equations 1.4-1.6). The variation of particle density and mean particle diameter can also cause minimum fluidization voidage to change by about 5% which is evaluated from the theoretical variation of Umf using Ergun's equation (1.1). The particle size distribution was found to affect Umf- The wide distribution has significantly lower Umf (7-13%) than the bimodal distribution and the intermediate fraction with the latter two having roughly the same Umf (less than 3% difference). The minimum fluidization voidages for all three size distributions differ by less than 4.5% which is not significant. Static bed height is not a factor in the determination of minimum fluidization velocity. When the bed height to diameter ratio changes from 3.3 to about 4.4, there is no definite shift in the Umf of F C C powders and the difference in Umf is mostly less than 3%. Frantz (1966) claimed that the bed height does not affect Umf as long as the bed height to diameter ratio is larger than two. In a packed bed, the bed pressure drop increases as superficial velocity increases. At minimum fluidization, the weight of the bed should be completely supported by the air flow. Table 4.1 shows that the bed pressure drop at minimum fluidization is within 2% of the predicted value for F C C . Static bed height is not a factor. Generally, a fluidized bed is regarded to having uniform gas distribution when the experimental bed pressure drop is within 5% of the pressure exerted by the weight of powders on the distributor Chapter 4. Results and Discussion: Low Velocity Fluidization 51 Table 4.1: Minimum Fluidization Data for Fluid Cracking Catalyst. Static bed heights are 3.3 and 4.4 times the bed diameter Size Distribution H0 = 3.3D H0 = 4AD Experimental um/ (m/s) exp'lAPmf pred. APmf Experimental um/ (m/s) exp'lAPmf pred.APm, Original 0.00208 0.99 0.00214 0.99 Coarse 0.00553 1.00 0.00585 0.99 Intermediate 0.00285 0.99 0.00275 0.98 Fine 0.00151 1.00 0.00153 1.00 Wide 0.00250 0.99 0.00257 0.99 Bimodal' 0.00287 0.99 0.00282 0.99 Chapter 4. Results and Discussion: Low Velocity Fluidization 52 Table 4.2: Experimental Values of Minimum Fluidization Voidage for F C C at Different Static Bed Heights Size Distribution Experimental em/ HQ = 3.3D H0 = 4AD Difference due to Change in H Original 0.497 0.491 -1.2% Coarse 0.478 0.473 -1.0% Intermediate 0.501 0.497 -0.8% Fine 0.554 0.546 -1.4% Wide 0.479 0.475 -0.8% Bimodal 0.483 0.483 0.0% Chapter 4. Results and Discussion: Low Velocity Fluidization 53 plate. No channels were visible during fluidization of F C C powders. Bubbles were fairly uniformly distributed as they reached the bed surface. The cardboard distributor plate provided better gas distribution than other types of distributor such as the metal screen or perforated plate. Among the correlations for Umf considered here, the one proposed by Baeyens and Gel-dart (1973) gave the best results. As shown in Table 4.3, Baeyens and Geldart's prediction is less than 20% different from the actual values of Umf- Other correlations produced predictions that differ from the experimental data by as much as 150%. All the correlations for minimum fluidization velocity predict that Umf is proportional to about the square of the mean particle diameter for small particles. When all the size distributions of FCC used in the experiment are included, the following relationship is obtained: Umf cc d)f (4.1) If only the powders with larger mean particle size (i.e. coarse, intermediate, original, bimodal and wide distributions) are considered, then: Umf CC d)f ' (4.2) This indicates that the dependence of minimum fluidization velocity is reduced for smaller particles, similar to the findings of Simone and Harriott (1980) who explained that the large deviation for the smallest size was largely due to the higher minimum fluidization voidage for the fine powder (Table 4.2). A low dependence of minimum fluidization velocity on the surface-volume mean par-ticle diameter has also been obtained by Frantz (1966). He argued that the deviation was caused by the use of powders with wide size distributions. Most of the correlations previously shown were developed for narrow size distribution powders. His argument is not substantiated by the results collected here. When only the coarse, intermediate and Chapter 4. Results and Discussion: Low Velocity Fluidization 54 Table 4.3: Predicted Minimum Fluidization Velocities for F C C Powders. Bracketted Values are the Ratio of Predicted to Experimental Values. Size Distribution Predicted Umf (m/s) Ergun Wen and Yu Baeyens and Davies and (1952) (1966) Geldart (1973) Richardson (1966) Original 0.00178 0.00182 0.00191 0.00273 (0.85) (0.87) (0.91) (1.31) Coarse 0.00584 0.0133 0.00554 0.00623 (1.06) (2.43) (0.97) (1.10) Intermediate 0.00250 0.00502 0.00259 0.00267 (0.87) (1.77) (0.92) (0.95) Fine 0.00110 0.00167 0.00124 " 0.00117 (0.72) (1.10) (0.80) (0.75) Wide 0.00244 0.00278 0.00253 0.00260 (0.97) (1.11) (1.01) (1.02) Bimodal 0.00255 0.00318 0.00264 0.00272 (0.88) (0.87) (1.04) (1.07) Chapter 4. Results and Discussion: Low Velocity Fluidization 55 fine fractions are considered, then: ct d, 1.45 (4.3) The index actually drops when only the narrow size distributions are considered. It should be recalled that the particle density of FCC changes from one size fraction to another. This could also contribute to the difference between the theoretical and experimental degrees of dependence. 4.1.2 Glass Beads The minimum fluidization properties for glass beads are shown in Table 4.4. The min-imum fluidization points were also very easy to identify from the plots of bed pressure drop versus superficial velocity. With decreasing mean particle diameter, Umf decreased and Cm/ increased. emf is generally smaller for glass beads than for the corresponding F C C fractions. The particle size distribution of the glass beads had a significant influ-ence on Umf. Among the intermediate, wide and bimodal distributions, the difference in surface-volume mean diameter was 1.7% which can theoretically cause 3.4% change in Umf and 5% in tmf- However, the difference in Umf was as much as 68%, with the intermediate fraction having the highest and the bimodal distribution the lowest values of Umf. The minimum fluidization voidage follows the same pattern. These are very significant differences, and they cannot be accounted for by the small difference in mean particle diameter. The uniformity of the gas distribution in the fluidized bed was more of a problem for the glass beads. The difference between the experimental bed pressure drop at minimum fluidization and the predicted value was as much as 5% which was still acceptable. Some channels, 30 mm long at most, were occasionally visible when the glass beads were fluidized. The bubble distribution was fairly uniform on the bed surface for the bubbling Chapter 4. Results and Discussion: Low Velocity Fluidization 56 Table 4.4: Experimental Minimum Fluidization Data for Glass Beads Size Distribution <W (m/s) prtd.APm, Coarse 0.442 0.0118 0.99 Intermediate 0.479 0.00588 0.99 Fines 0.495 0.00164 0.95 Wide 0.444 0.00501 1.00 Bimodal 0.406 0.00350 1.00 Chapter 4. Results and Discussion: Low Velocity Fluidization 57 regime. Hence the gas distribution was satisfactory for the fluidization of glass beads. The difference between the two types of material studied is the low tmf of the glass beads bed compared to the F C C . The particle density of the glass beads is also about 80% higher than that of the F C C . The glass bead fractions also showed some stickiness as observed by Khoe (1988), making the powder harder to fluidize and more prone to channelling. The correlations for Umf do not do as well for the glass beads as for F C C . As shown in Table 4.5, all but one of the predictions are at least 28% off the experimental values. None of the correlations provides satisfactory predictions for all distributions of glass beads. The dependence of Umf on surface-volume mean particle diameter is significantly higher for the glass beads than for the F C C . The following relationship is obtained when all the size distributions of glass beads are considered: Umf oc d)v75 (4.4) There is no significant difference when only the narrow size fractions (coarse, intermediate and fine) are included: Umf a dlJ* (4.5) The index goes up to beyond 2 when the size fraction with the smallest mean particle diameter is excluded: Umf « d]f (4.6) Simone and Harriott's explanation (1980) that the smallest particles are mainly respon-sible for lowering the dependence of Umf on the surface-volume mean diameter appears to be correct for glass beads. It should be pointed out that the lowest mean particle size among the glass bead fractions was 71 pm when the fine distribution is excluded while Chapter 4. Results and Discussion: Low Velocity Fluidization 58 Table 4.5: Predicted Minimum Fluidization Velocities for Glass Bead Powders. Bracket-ted values are the ratio of predicted to experimental values. Size Distributions Predicted Umf (m/s) Ergun Wen and Yu Baeyens and Davies and (1952) (1966) Geldart (1973) Richardson (1966) Coarse 0.0190 0.0200 0.0164 0.0203 (1.63) (1.72) (1.41) (1.75) Intermediate 0.00787 0.00568 0.00740 0.00840 (1.37) (0.98) (1.28) (1.46) Fine 0.00208 0.00084 0.00224 0.00222 (1.28) (0.51) (1.38) < (1.37) Wide 0.00771 0.00622 0.00725 0.00821 (1.56) (0.65) (1.46) (1.66) Bimodal 0.00795 0.00643 0.00747 0.00849 (2.33) (1.87) (2.18) (2.49) Chapter 4. Results and Discussion: Low Velocity Fluidization 59 it was 53 pm for the F C C . Even the intermediate fraction of FCC is probably too fine to have the expected dependence of Umf on the mean particle size. 4.2 Bubble-Free Bed Expansion 4.2.1 Fluid Cracking Catalyst Bubble-free bed expansion data are shown in Table 4.6. The correlations provided by Abrahamsen and Geldart (1980a) (equation 1.9) indicated that decreasing the mean particle size should increase the maximum bubble-free expansion, e^, of a batch of powder. The narrow size distribution fractions are considered first. Going from the coarse (d,v = 83^m) to the intermediate (d,v = 53pm) to the fine (dtv = 34pm) fractions, the maximum bubble-free voidage increases by 10% and 12% respectively. The maximum variation of maximum bubble-free bed voidage when experiments were replicated was ±0 .4% of the mean. Hence the effect of mean particle diameter on emb is very significant. Changing the static bed height does not seem to affect the maximum bubble-free bed voidage significantly. As mentioned in Section 1.2.2, the bubble-free bed expansion may depend on the elasticity coefficient which in turn depends on the number of contact points among par-ticles. For the same volume of powder, the number of contact points for the fine fraction is obviously more than for the coarse fraction. Therefore the powder with a smaller surface-volume mean particle size expands more than that with a larger mean particle size. Mutsers and Rietema (1977) mentioned that powder of wide size distribution has more contact points per unit volume than a powder with a narrow size distribution. Hence the wide distribution powder is expected to expand more homogenously than the intermediate distribution. In our experiments, while the maximum bubble-free voidages Chapter 4. Results and Discussion: Low Velocity Fluidization 60 Table 4.6: Bed Voidages at Minimum Bubbling and Corresponding Fractional Bub-ble-Free Bed Expansion Data for Different F C C Size Distributions at Different Static Bed Heights. Size Distribution Minimum Bubbling Voidage Fractional Bed Expansion H„ = 3.3£> H0 = 4.41? H0 = 3.3£> H0 = 4 AD Original 0.558 0.550 0.123 0.120 Coarse 0.505 0.500 0.056 0.057 Intermediate 0.549 0.548 0.096 0.100 Fine 0.617 0.613 0.114 0.123 Wide 0.528 0.530 0.102 0.116 Bimodal 0.528 0.529 0.093 0.095 Chapter 4. Results and Discussion: Low Velocity Fluidization 61 (emft) of the wide and bimodal F C C powders differed by only 0.4% from each other, they were about 3.6% less than that of the intermediate distribution. However, these three distributions have slightly different mean particle diameters (1.7%) and particle densities (2.2%) which, in theory (estimated using equation 1.12), can result in a 3% change in emfc. Since the variation of em(, determined in replicated measurements was 0.4%, the difference in has to be at least 3.5% in order to be significant. Hence, increasing the spread of particle size distribution significantly lowers the minimum bubbling voidage. The wide distribution actually had significantly higher fractional bed expansion (6-16%) than the intermediate F C C powder. The ratio of interparticle void volume to bulk volume of solids is normally higher for a powder with narrow size distribution. Even though the fractional bed expansion may be larger for the wide distribution powder, it may not be enough to make up the difference in static bed voidage so that the wide distribution has a higher maximum bubble-free voidage than the intermediate fraction. The bimodal powder contains 67.5% by weight of coarse fraction and 32.5% of fine fraction. Compared to the wide F C C distribution, the bimodal distribution has higher fines fraction which promotes bubble-free bed expansion. On the other hand, the higher coarse content in the bimodal distribution powder discourages such expansion. The overall result is that the coarse powder predominates and the bimodally distributed powder has significantly lower fractional bubble-free expansion when compared with the wide fraction of F C C . The wide distribution powder expands 4-6% less than the original F C C powder. The principal difference between these two powders is the higher coarse particle content in the wide distribution. Bubble-free expansion is reduced with the addition of coarse powder. Of course, one has to be careful in comparing these two powders because the wide distribution has a somewhat higher mean particle diameter than the original material. Chapter 4. Results and Discussion: Low Velocity Fluidization 62 4.2.2 Glass Beads With the glass beads, the minimum bubbling voidage and the fractional bubble-free bed expansion again increase as the mean particle size decreases (Table 4.7). Though the coarse fraction falls barely into the Geldart's Group B powder category, it still expands to a small but significant extent. This simply shows that the change in fluidization behavior from Group A to Group B powders is gradual rather than stepwise. The difference in maximum bubble-free bed expansion between the coarse and the intermediate fractions is much less than that for FCC. This trend persists when the intermediate and the fine fractions are compared. This may be caused by the change in particle density, tmf and Umf. In general, the mean particle sizes of the corresponding FCC fractions are smaller than those of the respective fractions of glass beads. It is therefore natural that the former would expand more before reaching Umb than the latter. However, it should also be noted that the intermediate and fine fractions of the glass beads appeared to be quite sticky. The interparticle forces for these tractions may be greater than the forces the fluid can exert on the particles. Hence some channelling occurs when fluidizing these glass bead fractions and the powders do not expand as much as expected. The intermediate, bimodal and wide distributions differ in mean particle diameter by 1.7% which can theoretically cause a deviation with the same magnitude in emb among the three powders. This plus the experimental variations from replicated measurements imply that the difference in emb for these size distributions must be at least 2.5% in order to be significant. The intermediate powder was found to have a significantly higher maximum bubble-free bed expansion (3.7%) than the wide distribution though the latter had a higher fractional bed expansion as predicted by Mutsers and Rietema (1977). Again this is due to the much higher Emf of the intermediate compared to the wide distribution powders. The bimodal distribution had a emb value 3% less than for the wide Chapter 4. Results and Discussion: Low Velocity Fluidization 63 Table 4.7: Bed Voidage at Minimum Bubbling and Corresponding Fractional Bubble-Free Bed Expansion for Glass Bead Powders. Size Distribution Minimum Bubbling Voidage Fractional Bed Expansion * C m / ' Coarse 0.442 0.452 0.023 Intermediate 0.479 0.492 0.027 Fine 0.495 0.516 . 0.042 Wide 0.444 . 0.474 0A.068 Bimodal 0.406 0.460 0.133 Chapter 4. Results and Discussion: Low Velocity Fluidization 64 distribution, and this difference is also significant. However, the bimodally distributed glass beads expanded fractionally about 100% more than the wide distribution and that is different from the results for F C C . The relative trend of fractional bed expansion between the wide and narrow size distributions appears to be generally true for Group A powders. However, the fractional bubble free expansion of the bimodal distribution relative to the other fractions should be treated cautiously. The bimodal distribution is made up of two other size fractions whose proportions change according to the desired mean particle diameter. For some systems, the coarse particles predominate as in the case of F C C and the overall result is that the bed expands fractionally less than a wide and continuous size distribution powder with the same mean particle diameter. For other systems, the fine fraction predominates as in the case of glass beads, and the mixture expands fractionally more than the corresponding wide size distribution powder. The order of maximum bubble-free bed voidage (em(,) for different size distributions also should not be generalized because it depends very much on the difference in static bed voidage as well as fractional bed expansion among the various size distributions. Another interesting point is that both the wide and the bimodal distributions of glass beads expand fractionally more than the fine fraction by a significant 50-200%. The mean particle diameters of the wide and bimodal distribution powders are almost twice that of the fine fraction so that one would expect the fine fraction to expand more. However, the interparticle forces are so strong in the fine fraction that channelling occurs. When a small amount of fines is added to a powder of much larger particle diameter, the fines fill up the interparticle spaces between the larger particles. The overall attraction forces among the particles in this powder are then stronger than among big particles only, but weaker than among fine particles only. This makes the powder less sticky than the fine fraction so that channelling is less likely to occur. The powder then has a chance to expand evenly. The stronger attraction forces of the wide size spectrum powder enable Chapter 4. Results and Discussion: Low Velocity Fluidization 65 a higher degree of expansion compared to a powder of similar mean particle size but narrow size spectrum. 4.3 M i n i m u m Bubbling Properties 4.3.1 F lu id Cracking Catalyst Minimum bubbling velocities for F C C are shown in Table 4.8. The minimum bubbling points for the F C C were very easy to identify. At gas velocities just below Umbi numerous tiny air jets could be seen on the bed surface resembling the appearance of volcanoes. This must not be mistaken as the minimum bubbling point. At the minimum bubbling point, gross bubbling appeared at three or more areas on the bed surface. The bubble distribution was uniform and widespread with no gas maldistribution. The maximum fluctuations in Umb and emb determined in replicated measurements were ± 3 % and ±0.4% of their respective means. As the mean particle diameter decreased from 83/im (coasrse fraction) to 34 i^m (fine fraction), the minimum bubbling velocity decreased by 62%. However, the minimum bubbling voidage (see Table 4.6) was found to increase by 22%. The minimum bubbling properties were affected by the particle size distribution. Theoretically, the differences in mean particle size and particle density among the in-termediate, wide and bimodal distributions can cause changes of 2% in Umb and 3% in m^b (estimated using equations 1.10 and 1.12). That means the difference in Umb has to be over 5% in order to be significant since the maximum variation of Umb was 3% when experiments with the same powder were replicated. The bimodal distribution has significantly higher Umb than the intermediate fraction (9-11% difference). However, the Umb of the wide distribution does not differ significantly from either the intermediate (narrow) or the bimodal distributions. It was pointed out in Section 4.2 that increasing Chapter 4. Results and Discussion: Low Velocity Fluidization 66 Table 4.8: Experimental Minimum Bubbling Properties for Different F C C Size Distribu-tions. Size Distribution Umb(m/s) Effect of Increase in Static Bed Height H0 = 3.3£> H0 = 4AD A t / m 6 Original 0.00456 0.00449 -1.6% -1.4% Coarse 0.00890 0.00857 -3.7% -1.0% Intermediate 0.00503 0.00514 2.2% -0.2% Fine 0.00371 0.00383 3.2% -l.oVo Wide 0.00531 0.00533 0.4% 0.4% Bimodal 0.00550 0.00570 3.6% 0.2% Chapter 4. Results and Discussion: Low Velocity Fluidization 67 the spread of particle size distribution significantly lowers the minimum bubbling voidage of F C C powders. Changing the static bed height does not seem to have a definite effect on the minimum bubbling velocity or minimum bubbling voidage. Except for the fluidization of Group C powders, increasing the ratio of Umb/Umf is generally considered to result in an improvement in the quality of fluidization, mean-ing smaller pressure fluctuations and smaller gas bubbles. Larger ratios signify more importance of the bubble-free regime. Table 4.9 shows that the velocity ratio is well over one for all the F C C powders. When the mean particle size decreases, the quality of fluidization becomes better. Par-ticle size distribution also affects the quality of fluidization. It is normally expected that the addition of fines results in smoother fluidization by reducing gas maldistribution and enhancing dense phase gas flow . The velocity ratio for the wide size distribution is significantly higher (11-20%) than for the intermediate (narrow cut) distribution. Cor-respondingly, the fluidization is generally better in powders with wide size spectra than in those of narrow particle size spread. There is no significant difference in Umb/Umf be-tween the bimodal distribution and either the intermediate or the wide distributions even though the the bimodal distribution has a distinctly higher Umb than the intermediate fraction. The predicted minimum bubbling properties are shown in Table 4.10. The simpler correlation for Umb (equation 1.10, Geldart and Abrahamsen (1978)) gives much better predictions than the more complicated (equation 1.11, Abrahamsen and Geldart (1980a)). Ironically, the latter which takes the fines fraction into consideration produces the worst prediction for the fine fraction of the F C C . These results do not support Abrahamsen and Geldart's (1980a) belief that the more extensive correlation should be used when the weight fraction of powder with particle diameter of less than 45 pm exceeds 15%. In general, the simple correlation predicted Umb to within 10% of the actual value. Foscolo's Chapter 4. Results and Discussion: Low Velocity Fluidization 68 Table 4.9: Ratios of Minimum Bubbling Velocities to Minimum Fluidization Velocities for Different F C C Size Distributions. Size Distribution Original Coarse Intermediate Fine Wide Bimodal u. m/ H0 = 3.3D 2.19 1.61 1.76 2.46 2.12 1.92 H0 = 4.4D 2.10 1.46 1.87 2.50 2.07 2.02 Chapter 4. Results and Discussion: Low Velocity Fluidization 69 Table 4.10: Predicted Values of Minimum Bubbling Velocities and Voidages for Differ-ent Size Distributions of F C C . Bracketted values are the ratios of the predicted to the experimental values. Size Distribution F22 F45 (m/s) Geldart and Abrahamsen (1978) Abrahamsen and Geldart (1980a) Foscolo (1983) Original 0.0087 0.357 0.00450 (0.99) 0.00619 (1.37) 0.507 (0.92) Coarse 0.0000 0.0124 0.00833 (0.96) 0.00887 (1.02) 0.393 (0.71) Intermediate 0.0000 0.0992 0.00532 (1.05) 0.00612 (1.21) < 0.465 (0.85) Fine 0.0031 0.885 0.00341 (0.90) 0.00694 (1.85) 0.572 (0.93) Wide 0.0050 0.188 0.00531 (LOO) 0.00652 (1.22) 0.470 (0.91) Bimodal 0.0003 0.294 0.00540 (0.97) 0.00712 (1.28) 0.463 (0.88) Chapter 4. Results and Discussion: Low Velocity Fluidization 70 (1983) correlation for emt (equation 1.12) worked well for powders of small mean particle diameter giving predictions within 15% of the true values. For the coarse fraction, the predictions are out by as much as 29%. 4.3.2 Glass Beads The minimum bubbling results for the glass beads are given in Table 4.11. The minimum bubbling points for most size distributions were again quite distinct except for the fine fraction where localized bubbles appeared at gas velocities just beyond Umf- The strong interparticle forces between these fine particles makes bubble-free bed expansion uneven. The values of minimum bubbling velocity and voidage given here correspond to where bubbles were seen over most of the bed surface. This involved some subjective judgement so that the results for this traction should be treated with caution. As for F C C , the minimum bubbling velocity of the glass beads goes down while the minimum bubbling voidage (see Table 4.7) goes up with decreasing mean particle size. The theoretical variations of Umb caused by small difference in mean particle diameter among the intermediate, wide and bimodal distributions are less than 1.7%. For the distributions with similar mean particle diameter, the minimum bubbling velocities of the bimodal and wide distributions differ by 2.2% which is not significant because it has the same magnitude as the maximum fluctuations oiUmb determined in replicated measurements. However, the bimodal and wide distributions have minimum bubbling velocities of 13-15% lower than the intermediate (narrow) fraction and that is significant. This differs from the results for F C C . The minimum bubbling velocity of wide F C C powder is higher than that of the corresponding intermediate fraction. As described in Section 1.2.3, the effects of particle size distribution on the Umb is controversial. It seems that Umb may also depend on other particle properties such as particle shape and particle density. Some particles are porous while others are not. The minimum bubbling voidage Chapter 4. Results and Discussion: Low Velocity Fluidization 71 Table 4.11: Experimental Minimum Bubbling Data Glass Beads. Size Distribution umb (m/s) Urn* Coarse 0.0141 1.19 Intermediate 0.00883 1.50 Fine 0.00572 3.49 Wide 0.00769 1.53 Bimodal 0.00752 2.15 Chapter 4. Results and Discussion: Low Velocity Fluidization 72 may also be a factor. As shown in Section 4.2.2, the wide and bimodal distributions also have significantly lower emb than the intermediate fraction. The trends in the Umb/Umf ratio are also different for both materials. Table 4.11 shows that the intermediate (narrow) and the wide distributions of glass beads have essentially the same Umb/Umf ratio (2% difference). The highest ratio occurred for the bimodal distribution of the glass beads rather than the wide distribution as for the F C C . According to Abrahamsen and Geldart (1980b), the Umb/Umf ratio depends on the fines fraction, F 4 5 . Since the intermediate fraction has fewer fines than the other distributions, it is expected that it should also have a low Umb/Umf ratio. From this, one would expect a low Umb for the intermediate fraction. However, the intermediate fraction also has the highest Umf and this more than compensates for the effect of the low level of fines. This produces a high Umb f ° r the intermediate fraction. The bimodal distribution has a higher Umb/Umf ratio than the wide distribution. This suggests that the fines portion of the bimodally distributed glass beads prevails. The predicted minimum bubbling properties of the glass beads (see Table 4.12) are generally not as good as for the F C C (see Table 4.10). The simpler correlation (equa-tion 1.10) for Umb provides predictions that differ by as much as 35% from the experi-mental values. Foscolo's (1980) predictions of tmb with equation (1.12) are in error by as much as 48%. However, the correlation that takes the fines fraction into consideration provides much better predictions for glass beads than for F C C . The maximum difference between the experimental and predicted values is only 21%. For glass beads with narrow size distribution, this correlation is better than the simple correlation for Umb-Chapter 4. Results and Discussion: Low Velocity Fluidization 73 Table 4.12: Predicted Minimum Bubbling velocities and Voidages for Different Size Distributions of Glass Beads. Bracketted values are the ratios of the predicted to the experimental values. Size Distribution F22 F 4 5 umb {mfs) €mb Geldart and Abrahamsen (1978) Abrahamsen and Geldart (1980a) Foscolo and Gibilaro (1983) Coarse 0.0000 0.000 0.0113 (0.81) 0.0120 (0.86) 0.233 (0.52) Intermediate 0.0001 0.017 0.00727 (0.84) 0.00784 (0.90) 0.319 (0.65) Fine 0.102 0.587 0.00374 (0.65) 0.00606 (1.07) 0.466 (0.90) Wide 0.0114 0.090 0.00719 (0.95) 0.00888 (1.18) 0.321 (0.68) Bimodal 0.0300 0.184 0.00731 (0.99) 0.00888 (1.21) 0.318 (0.69) Chapter 4. Results and Discussion: Low Velocity Fluidization 74 4.4 Summary With increasing surface-volume mean particle size, Umf increases and emf decreases. Increasing the spread of particle size distribution with constant mean particle diameter decreases the minimum fluidization voidage and velocity for glass beads. The wide F C C powder has lower Umf than the bimodal and intermediate distributions although all three distributions have similar em/. Static height does not affect the minimum fluidization properties. The dependence of Umj on the mean particle diameter is stronger for coarser powders (FCC and glass beads). The correlation of Baeyens and Geldart (1973) for Um} is the best for F C C powders. Both the minimum bubbling voidage and the fractional bubble-free bed expansion increase with decreasing mean particle size. With the mean particle size kept constant, the wide distribution powder has higher fractional bed expansion than with narrow cut (intermediate) fraction although the latter has a higher minimum bubbling voidage. Different materials have different minimum bubbling properties. Decreasing the mean particle size decreases the minimum bubbling velocity, increases the minimum bubbling voidage and leads to smoother fluidization unless the powder becomes too cohesive. The Umb depends on factors other than the mean particle size and particle size distribution. Increasing the spread of particle size distribution while maintaining a constant mean particle size decreases emb of both glass beads and F C C . Chapter 5 Results and Discussion: Dense Phase Properties As mentioned previously, collapse tests were performed to measure the dense phase prop-erties of some powder fractions. A n example of bed height versus time plot from the col-lapse test of a powder studied in our experiment is shown in Figure 5.1. The linear part in the middle denotes the 'hindered sedimentation stage'. The dense phase bed height is the intercept of this linear portion and the ordinate axis. The superficial collapse velocity is the slope of the linear section of the collapse curve. Since the air pressure above and below the distributor plate are equalized throughout the course of bed collapse, no gas could flow through the distributor in either direction. Thus the bed collapse velocity is essentially the superficial dense phase gas velocity (Geldart and Wong, 1984). The results were anslyzed with confidence level of 95%. As an example, error bars are shown on the plots for the dense phase properties of F C C with the static bed height 4.4 times bed diameter. The de-aeration rate of the F C C powders was in general slower than the glass beads, i.e. the former has a longer period of hindered sedimentation. It is sometimes more diffi-cult to identify the linear section on the bed height versus time plot for some glass bead fractions. This creates more uncertainty in the determination of dense phase properties. Three to four runs were performed to minimize the experimental errors and improve the reliability of the data. 75 Chapter 5. Results and Discussion: Dense Phase Properties 76 Figure 5.1: Sample Plot of Bed Height Versus Time for Intermediate F C C Powder with Initial Superficial Velocity of 0.0086 m/s. Chapter 5. Results and Discussion: Dense Phase Properties 77 5.1 Effect of Mean Particle Diameter of Fluid Cracking Catalyst The collapse test results for the fluid cracking catalyst powder with different surface-volume mean particle diameters are shown in Figures 5.2-5.5. The dense phase superficial velocity and voidage at minimum fluidization are given on the plots as reference points. For all the size distributions of F C C , the dense phase voidage increases as expected when the superficial velocity is increased from Umf to Umb- During this stage, gas flows through the dense phase only. At about the minimum bubbling point, both the superficial dense phase gas velocity and the dense phase voidage reach their peaks. The interparticle gaps are at their largest. When the superficial velocity is increased beyond this point, bubble-free expansion is no longer possible and bubble formation occurs. Gas is now carried and passed through the bubble phase. This means a drop in superficial dense phase gas velocity. Since the dense phase voidage is related to the dense phase gas flow, both the dense phase gas flow and voidage keep on dropping. Eventually, the values of these two dense phase properties level off and reach limiting values. The amount of dense phase contraction in response to a certain increase in gas velocity depends on the mean particle size when the narrow size fractions (coarse, intermediate and fine) are considered. The extent of contraction increases with decreasing mean particle size. The limiting dense phase voidage and gas flow are also functions of mean particle size, both being higher for larger particles. The superficial gas velocity was increased to about 38 mm/s which is well beyond Umb f ° r a n y s i z e fraction. At this flowrate, the dense phase gas velocity and bed voidage hardly changed with increasing gas flow. Therefore the dense phase properties measured at U = 38 mm/s can serve as good estimations for much higher superficial velocities (in the bubbling regime) at which most industrial processes are operated. The dense phase voidage at high gas flows (Table 5.1) is roughly the same as the Chapter 5. Results and Discussion: Dense Phase Properties 78 ID O ( ) O IT) ^* ° i 1 1 1 1 r 1 r— 1 " 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 GRS VELOCITY (U - Umf) (m/s) Figure 5.2: Dense Phase Voidages for FCC Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 3.3 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 79 o <=>-. i n ID o . o o ID o IT) o jb CO \ o - £ 8. 'o d , . IT) o . >- • ID O o o o J o 13 D O A LEGEND ORIGINAL S I Z E DISTRIBUTION COARSE S I Z E DISTRIBUTION INTERMEDIATE S I Z E DISTRIBUTION S I Z E DISTRIBUTION 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 GAS VELOCITY (U - Umf) (m/s) 0.040 0.045 Figure 5.3: Bed Collapse Velocities for F C C Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 3.3 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 80 t o CO o ~i r 0.000 0.005 0.010 0.015 0.020 GRS VELOCITY (U 0.025 0.030 0.035 - Umf) (m/s) 0.040 0.045 Figure 5.4: Dense Phase Voidages for FCC Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 4.4 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 81 o o o in co o . o o to o in in o . , o CO _ £ i O X o in o . f—< in ^ • LEGEND ORIGINAL S I Z E DISTRIBUTION CORRSE S I Z E DISTRIBUTION INTERMEDIATE S I Z E DISTRIBUTION PINE S I Z E DISTRIBUTION 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 GAS VELOCITY (U - Umf) (m/s) 0.040 0.045 Figure 5.5: Bed Collapse Velocities for F C C Powders with Different Surface-Volume Mean Particle Diameters. Static bed height is 4.4 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 82 Table 5.1: Dense Phase Properties in Vigorously Bubbling Beds for Different Size Dis-tributions of F C C . Size Distribution ud (*10-3m/s) ed H0 = 3.3D H0 = AAD H0 = 3.3D H0 = 4AD Original 2.81 2.40 0.503 0.494 Coarse 5.95 5.80 0.485 0.468 Intermediate 3.24 2.80 0.507 0.501 Fine 1.85 1.85 0.566 0.557 Wide 3.45 3.12 0.488 0.478 Bimodal 3.55 3.25 0.492 0.482 Chapter 5. Results and Discussion: Dense Phase Properties 83 emf for all the F C C size distributions. This probably implies that the powder has little elasticit)r, regardless of the mean particle size. The superficial dense phase gas velocity increases and the dense phase voidage decreases for powders with smaller mean particle size for both of the bed heights considered. At first sight, it may seem more reasonable to use large particles in a chemical reactor to take advantage of the higher dense phase gas flow. However, larger particles mean smaller surface areas or effectiveness factors of the particles, slowing down the reaction. Hence it is important to increase the dense phase gas flowrate while keeping the particle size small. One way of doing this is by altering the particle size distribution. 5.2 Effect of Particle Size Distribution for Fluid Cracking Catalyst Collapse tests for three size distributions of F C C with similar mean particle size (interme-diate, wide and bimodal) are given in Figure 5.6-5.9. The intermediate size fraction has the smallest capacity to expand during bubble-free bed expansion. The dense phase of the bimodal and wide distributions (both with broader size spectra) are more capable of expanding. Though the intermediate fraction still has the highest peak e<j, the difference between the different size distributions is only a quarter of the difference at minimum fluidization (see Table 4.2). The difference in Umf among the three size distributions is about 0.3 mm/s while the peak dense phase velocity is about 1 mm/s higher for the wide distribution compared to the intermediate fraction; that of the bimodal distribution is another 0.3 mm/s higher. The slope of the Uj versus U — Umf plot during the bubble-free fluidization stage is the same for all three size distributions. The intermediate fraction obviously cannot accommodate the increase in gas flow as much as the other broader size distributions. The superficial dense phase velocit3r of the intermediate fraction starts off at a value lower Chapter 5. Results and Discussion: Dense Phase Properties 84 ID ID LEGEND A = INTERMEDIATE S I Z E DISTRIBUTION o = B1M0DRL S I Z E DISTRIBUTION v = WIDE S I Z E DISTRIBUTION 1 1 1 1 1 1 1 — 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 GRS VELOCITY ( U - Umf ) (m/s) — I 0.040 0.045 Figure 5.6: Dense Phase Voidages for F C C Powders with Similar Mean Particle Diameters but Different Size Distributions. Static bed height is 3.3 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 85 o ID o o ID ID O _ D O LO O _  (0 £ ' o LO X t i o >-f — •——1 o o o _J O _ L i J o > u C O Q_ LO cn to _! o _ _J O O o C D L J C D O to O _ O LO CM o o o o a J LEGEND A = INTERMEDIATE SIZE DISTRIBUTION o = BIMODAL SIZE DISTRIBUTION v = WIDE SIZE DISTRIBUTION 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 GRS VELOCITY (U - Umf) (m/s) 0.040 0.045 Figure 5.7: Bed Collapse Velocities for FCC Powders with Similar Mean Particle Diam-eters but Different Size Distributions. Static bed height is 3.3 times bed diameter Chapter 5. Results and Discussion: Dense Phase Properties 86 in o O-OOO 0.005 0.010 0.015 0.020 0.025 0.030 0.035 GflS VELOCITY (U - Umf) (m/s) 1 I 0.040 0.045 Figure 5.8: Dense Phase Voidages for FCC Powders with Similar Mean Particle Diameters but Different Size Distributions. Static bed height is 4.4 times bed diameter. Error bars are the spread for 95% confidence level. Chapter 5. Results and Discussion: Dense Phase Properties 87 o U D O in in o o in o (0 O in - , TT x .-I >-O o L J o UJ cn CC to —1 c O CJ CD CO S o , in o C M o LEGEND INTERMEDIATE SIZE DISTRIBUTION BIMODAL SIZE DISTRIBUTION WIDE SIZE DISTRIBUTION 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 GAS VELOCITY (U - Umf) (m/s) 0.040 0.045 Figure 5.9: Bed Collapse Velocities for F C C Powders with Similar Mean Particle Diam-eters but Different Size Distributions. Static bed height is 4.4 times bed diameter. Error bars are the spread for 95% confidence level. Chapter 5. Results and Discussion: Dense Phase Properties 88 than its minimum fluidization velocity as found also by Rowe and Yacono (1976) for a size distribution with very few fines. At gas velocities roughly beyond Umb, the dense phase contracts and the superficial dense phase gas velocity decreases as with other size distributions of F C C . All three size distributions-follow the same pattern. The limiting is significantly higher for the intermediate fraction while those of the other two distributions are quite similar. In fact, the difference in ed between the intermediate fraction and the distributions with wider size spectra is re-established to its level at minimum fluidization. Hence it would appear that dense phase expansion plays little or no role in a vigorously bubbling bed. However, higher cj does not alwaj's mean higher dense phase gas flow. The perme-ability of the dense phase is affected by the particle size distribution of the powder. As shown in Figures 5.7 and 5.9, Ud is highest for the bimodal distribution and lowest for the intermediate fraction. The difference in dense phase properties between the narrow size cut and the broad size powders is significant at the 95% confidence level. This is in the opposite order when compared with the e<f. The increase in dense phase perme-ability in response to having a wide size spectrum more than compensates for the small dense phase volume in the bimodal and wide distributions. Therefore, they are able to accommodate high dense phase gas flow. Less gas goes through the bubble phase. No clear conclusion can be drawn on the difference in dense phase properties between the bimodal and wide distributions. The dense phase properties did not show consistent and significant changes when going from a bimodal to a wide distribution F C C powder. Hence, dense phase properties cannot be predicted by the fines fraction as the bimodal distribution powder has higher Fi5 but lower F22 compared with the wide distribution. The overall size distribution has to be considered. Increasing the static bed height decreases both the dense phase voidage and the superficial dense phase gas velocity. This is true for all the size distributions studied Chapter 5. Results and Discussion: Dense Phase Properties 89 here. This is discussed more thoroughly in Section 5.5. 5.3 Effect of Particle Size Distribution for Glass Beads The collapse test results for the glass beads are shown in Figures 5.10 and 5.11. There is a large deviation in the em/ among the intermediate, bimodal and wide size distributions. The intermediate and the bimodal distributions have the highest and lowest e<f values respectively, even in a bubbling bed. However, the difference is reduced near £7mf>. This again shows that the dense phase of the powders with wider size spectra have a larger capacity to expand during bubble-free fluidization. This influence of size distribution disappears when there is vigorous bubbling, and the dense phase voidage is re-established roughly to their respective tmj. The results for the superficial dense phase velocity differ from those of the F C C . When the superficial velocity is increased, Ud for each distribution continues to rise during the bubble-free fluidization stage, dips to a limiting value during the bed contraction stage and then increases very gradually again. The largest difference between the dense phases of the two types of powder occurs during vigorous bubbUng as shown in Tables 5.2-5.3. The intermediate and the wide distributions have about the same superficial dense phase velocity. The superficial dense phase velocity of the bimodal distribution is as much as 25% lower, but Ud of the intermediate fraction is about 15% and 33% higher than for the wide and bimodal distributions, respectively, at minimum fluidization. This shrinks to about 1% and 25% at higher gas flow. The change in ed between the minimum fluidization and vigorous bubbling is similar for all three distributions. Hence the permeability of the dense phase for the bimodal and the wide distributions is higher than that of the intermediate fraction. The increase in dense phase permeability in the wide distribution is barely enough to compensate for the smaller when compared to the intermediate Chapter 5. Results and Discussion: Dense Phase Properties 90 Figure 5.10: Dense Phase Voidages for Glass Bead Powders with Similar Mean Particle Diameter but Different Size Distributions. Static bed height is 4.4 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 91 LO CO o o o o °H 1 I 1 1 1 1 1 1 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045 GflS VELOCITY (U - Umf) (m/s) Figure 5.11: Bed Collapse Velocities for Glass Bead Powders with Similar Mean Particle Diameter but Different Size Distributions. Static bed height is 4.4 times bed diameter. Chapter 5. Results and Discussion: Dense Phase Properties 92 Table 5.2: Experimental and Predicted Dense Phase Voidages in a Vigorously Bubbling Bed for Different Size Distributions of Glass Beads. Bracketted numbers are the ratio of the predicted to the experimetal values. Size Distribution Experimental Abrahamsen and Geldart (1980b) Dry et al. (1983) Richardson and Zaki (1954) Foscolo and Gibilaro (1983) Kmiec (1982) Intermediate 0.464 0.470 (1.01) 0.350 (0.75) 0.326 (0.70) 0.390 (0.84) 0.372 (0.80) Wide 0.431 0.440 (1.02) 0.360 (0.84) 0.328 (0.76) 0.391 (0.91) 0.373 (0.86) Bimodal 0.395 0.405 (1.03) 0.386 (0.98) 0.300 (0.76) 0.365 (0.92) 0.348 (0.88) Chapter 5. Results and Discussion: Dense Phase Properties 93 Table 5.3: Experimental and Predicted Superficial Dense Phase Gas Velocities in a Vigor-ously Bubbling Bed for Different Size Distributions of Glass Beads. Bracketted numbers are the ratio of the predicted to the experimetal values. Ud{*10-2m/s) Size Distribution Experimental Abrahamsen and Geldart (1980) Dry et al. (1983) Richardson and Zaki (1954) Foscolo and Gibilaro (1983) Kmiec (1982) Intermediate 5.05 4.77 (0.94) 4.26 (0.84) 20.6 (4.08) 11.6 (2.30) 14.3 (2.83) Wide 5.00 4.10 (0.82) 2.70 (0.54) 15.0 (3.00) , 8.02 (1.60) 9.96 (1.99) Bimodal 3.70 2.93 (0.79) 1.86 (0.50) 11.0 (2.97) 5.42 (1.46) 6.77 (1.83) Chapter 5. Results and Discussion: Dense Phase Properties 94 fraction. However, the difference in ed between the intermediate and bimodal distribution is too large (0.06) to be overcome by the larger dense phase permeability of the latter. The difference in Ud between the wide and bimodal distributions is maintained when going from Umf to vigorous bubbling stage. Hence the permeability of the two distributions is about the same and the difference in Ud is caused by the difference in e^ . 5.4 Examination of Correlations for Dense Phase Properties Correlations for ed and Ud were described in Section 1.3.2 . They have been compared with our experimental results. Only the correlations from Abrahamsen and Geldart (1980b), equations 1.16-17, take into'account the effect of bed height. The results from other correlations are compared with the experimental results averaged from the two bed heights studied. Some correlations have Uj and implicitly depending on each other. In these cases, the experimental value of one parameter is used to predict the other parameter. The results are shown in Tables 5.2 to 5.5. It can be seen that Abrahamsen's correlations predicts £<* and Ud within 13% and 21% of the experimental values, respectivel}-. This is the best agreement among the five correlations studied here, both for the F C C and the glass beads. Predictions from Dry et al. (1983) are all too low, probably because particle density is excluded in his correlations. Richardson and Zaki's (1954) predictions of Ud and differ from the experimental values by about 300% and 30%, respectively. A comparsion of the experimental and predicted values of the index n is given in Table 5.6. In general, Richardson and Zaki's predictions of the dense phase properties get better when the predicted index n is close to the experimental value. The correlations given by Foscolo and Gibilaro (1983), and Kmiec (1982) underesti-mate the dense phase voidage of the F C C and glass beads by about 10-20%. At the same Chapter 5. Results and Discussion: Dense Phase Properties 95 Table 5.4: Predicted Dense Phase Voidages in a Vigorously Bubbling Bed for Different Size Distributions of F C C . Bracketted numbers are the ratio of the predicted to the experimental values.  Size Distribution Abrahamsen and Geldart (1980b) Dry et al. (1983) Richardson and Zaki (1954) Foscolo and Gibilaro (1983) Kmiec (1982) H0 = 3.3D H0 = 4AD Original 0.561 (1.12) 0.534 (1.08) 0.357 (0.71) 0.418 (0.84) 0.446 (0.89) 0.441 (0.88) Coarse 0.504 (1.04) 0.493 (1.05) 0.344 (0.72) 0.386 (0.81) 0.429 (0.90) 0.409 (0.86) Intermediate 0.548 (1.08) 0.538 (1.07) 0.344 (0.68) 0.399 (0.79) 0.435 (0:78) 0.424 (0.75) Fine 0.636 (1.12) 0.626 (1.13) 0.348 (0.62) 0.437 (0.78) 0.458 (0.82) 0.454 (0.81) Wide 0.533 (1.09) 0.524 (1.10) 0.351 (0.73) 0.410 (0.85) 0.441 (0.91) 0.434 (0.90) Bimodal 0.539 (1.10) 0.533 (1.11) 0.344 (0.71) 0.407 (0.84) 0.440 (0.90) 0.433 (0.89) Chapter 5. Results and Discussion: Dense Phase Properties 96 Table 5.5: Predicted Superficial Dense Phase Gas Velocities in a Vigorously Bubbling Bed for Different Size Distributions of F C C . Bracketted numbers are the ratios of the predicted to the experimental values. l7d(*lC -3m/s) Size Distribution Abrahamsen and Geldart (1980b) Dry et al. (1983) Richardson and Zaki (1954) Foscolo and Gibilaro (1983) Kmiec (1982) H0 = 3.3D H0 = A AD Original 2.21 (0.79) 2.18 (0.91) 1.40 (0.54) 5.70 (2.18) 4.44 (1.70) 4.67 (1.79) Coarse 5.90 (0.99) 5.31 (0.92) 9.71 (1.67) 13.8 (2.36) 8.26 (1.41) 12.1 (2.06) Intermediate 2.99 (0.92) 2.82 (1.01) 3.66 (1.21) 8.41 (2.78) 6.37 (2.H) 6.87 (2.27) Fine 1.70 (0.92) 1.69 (0.91) 1.24 (0.67) 5.81 (3.14) 4.95 (2.68) 5.07 (2.74) Wide 2.71 (0.79) 2.60 (0.83) 2.10 (0.64) 6.77 (2.06) 5.07 (1.54) 5.47 (1.66) Bimodal 3.13 (0.88) 2.74 (0.84) 2.33 (0.69) 7.41 (2.18) 6.17 (1.81) 5.95 (1.75) Chapter 5. Results and Discussion: Dense Phase Properties 97 Table 5.6: Predicted Values of Index n, a Dense Phase Parameter, in a Vigorously Bubbling Bed for Different Size Distributions of F C C and Glass Beads. Bracketted numbers are the ratios of the predicted to the experimental values. index n Size Distributtion F C C Glass Beads Experimental Predicted Experimental Predicted Original 5.55 4.43 (0.80) - -Coarse 5.25 4.09 (0.78) - -Intermediate 5.86 4.37 (0.75) 5.81 3.98 < (0.69) Fine 6.49 4.53 (0.70) - -Wide 5.40 4.38 (0.81) 5.29 3.99 (0.75) Bimodal 5.45 4.37 (0.80) 5.14 3.97 (0.77) Chapter 5. Results and Discussion: Dense Phase Properties 98 time, they overestimate Ud by factors of two to three. 5.5 Effect of Static Bed Height on Dense Phase Properties The effects of static bed height on the dense phase properties for F C C are shown in Figures 5.12 and 5.13. The superficial dense phase velocity and dense phase voidage can be correlated with the static bed height by: ed a (5.1) Ud a HI (5.2) The values of /3 and 7 are shown in Table 5.7. One can see that the static bed height has a small effect on the dense phase voidage and a larger influence on the superficial dense phase gas velocity; in both cases, the dependence gets stronger with increasing superficial velocities. Pyle and Harrison (1967) postulated that there is a gradient of interstitial gas velocity along the height of the bed with the bubble phase underdeveloped in the region close to the distributor. Hence the interstitial velocity is high in this area. The local superficial dense phase gas velocitj' decreases with increasing height until it finally tapers off at certain bed level. The rate of this decrease of the superficial dense phase gas velocity varies from one type of powder to another. Rowe and Yacono (1976) drew a similar conclusion from their work. They used silicon carbide powder to make up beds varying from 0.2 to 0.6 m in depth with the mean particle sizes from 40 to 260 pm. They inferred that almost all gas flow occurred interstitially near the distributor. The permeability of the dense phase and the dense phase voidage decreased with increasing height while the interstitial gas flow approached Umf at the bed surface. Chapter 5. Results and Discussion: Dense Phase Properties 99 in LO LO to LO UJ CD cr o • — I o LJ Jo to - • cc ° X Q_ L J cn LJ a ™ o LO a - U o = U A = U + - U x = U o - U v - U B - U LEGEND - 2.8 mm/s = 4.3 mm/s = 4 . 6 mm/s = 8.8 mm/s 10.2 13.6 18.0 29.3 mm/s mm/s mm/s mm/s 0.2 0.3 0.4 0.5 STATIC BED HEIGHT (m) 0.6 0.7 Figure 5.12: Effect of Static Bed Heights on Dense Phase Voidages of Original FCC Distribution. Chapter 5. Results and Discussion: Dense Phase Properties 100 o LO O Ln o . d o „ d O LO X ° . . o >-> O o o ^ - J ° UJ O UJ CO cc <M _ i <=>• - J d o o Q L J ^ CD g in o o o o d • = U O = U A = U + - U X = U O - U v = U H = U LEGEND = 2.8 mm/s = 4.3 mm/s = 4.6 mm/s - 8.8 mm/s - 10.2 mm/s - 13.6 mm/s = 18.0 mm/s = 29.3 mm/s 0.2 0.3 0.4 0.5 STATIC BED HEIGHT 0.6 0.7 (ml Figure 5.13: Effect of Static Bed Heights on Bed Collapse Velocities of Original F C C Distribution. Chapter 5. Results and Discussion: Dense Phase Properties 101 Table 5.7: Indices for Variation of Dense Phase Voidage (6) and Superficial Dense Phase Velocity (7) on Static Bed Height at Different Gas Flowrates. u (*10-3m/s) P 7 2.80 -0.0299 -0.0140 4.30 -0.0204 -0.00375 4.60 -0.0155 -0.0384 8.80 -0.0513 -0.0638 10.2 -0.0543 -0.127 13.6 -0.0521 -0.123 18.0 -0.0754 -0.537 29.3 -0.0677 -0.614 Chapter 5. Results and Discussion: Dense Phase Properties 102 These past findings can perhaps help to explain the results from our experiments. During the bubble-free expansion stage, no bubble phase is present in the fluidized bed. There is only a small gradient in Ud caused by the hydrostatic pressure change along the bed. In the bed contraction stage, bubbles are present in the bed although not in large number. There is probably a certain gradient of Ud along the height of the bed. The powder retains some elasticity. An increase in the interstitial velocity brings about an increase in dense phase voidage. The gradient in ej at this stage is possibly larger than during bubble-free expansion. Hence the dependence of e<f and Ud on the static bed height is stronger. When the superficial velocity is increased further, the effect of the underdeveloped bubble phase at the lower portion of the bed may become more significant. The superficial dense phase velocity immediately above the distributor is quite high, and a velocity gradient may be present along a larger portion of the bed. The superficial dense phase velocity may not have levelled off even at the surface of some of the shallow beds used here. The dependence of the dense phase properties on the static bed height becomes even stronger. The degree of dependence in a vigorously bubbling bed is approximately as follows: ed cc H0-0 0 7 (5.3) Ud a tf-0-6 (5-4) The dependence of e<f on the static bed height is much lower than that of the superficial dense phase velocity. Hence only a small fraction of the decrease in Ud can be accounted for by the decrease in e<f with increasing static bed height. It is also known from the collapse test results that in a vigorously bubbling bed, the dense phase voidage stays at the limiting value in spite of the increase in superficial dense phase velocity with Chapter 5. Results and Discussion: Dense Phase Properties 103 increasing superficial velocity. Hence, the dependence of the dense phase voidage on the static bed height is relatively weak. The dependence of ej and the Ud on the static bed height are both two to three times higher than reported by Abrahamsen and Geldart (1980b) and Dry et al. (1983). Abrahamsen and Geldart used beds of about 0.3 to 0.9 m in depth while Dry et al. investigated beds more than 2 m deep. The bed depths in our experiments are only in the range of 0.28 to 0.67 m. The gradient in dense phase properties may have become negligible somewhere along the height of the deeper beds used by other workers. However, it is possible that the gradient is still present even at the surface of our more shallow beds. Therefore, a high degree of dependence is reported from our experiment. 5.6 Summary The dense phase properties depend on mean particle size, particle size distribution, su-perficial velocity, static bed height and physical properties of the powders. The dense phase of the powders with wide size spectra expands proportionally more during bubble-free fluidization than those with a narrow distribution. The dense phase voidage of any size distribution in a vigorously bubbling bed is roughly the same as the minimum flu-idization voidage; the latter is in turn controlled by the overall particle size distribution of the powder. The permeability of the dense phase is enhanced by having a broad size spectrum. The dense phase permeability and the dense phase voidage both affect the superficial dense phase gas velocity. The balance of these two factors determines which size distribution produces the highest dense phase gas flow. Decreasing the mean parti-cle size within Group A increases ed and decreases Ud- The effect of static bed height is stronger on Ud than on ej. Chapter 6 Results and Discussion: Pressure Fluctuations The final part of the experiments was intended to study pressure fluctuations in the bubbling and slugging flow regimes. The results were analyzed with 95% confidence level. The minimum slugging velocity for our powders is about 0.046 m/s (Stewart and Davidson, 1967). The fluidized bed should be in the bubbling regime for U = 0.037 m/s and in the slugging flow regime when U = 0.095, 0.175 and 0.267 m/s. Other criteria for the occurrence of slugging (Grace, 1982) were also met. The static bed height to diameter ratio was about five which is larger than the 3.5 required (Darton et al., 1977). The maximum stable bubble size for the powders studied is at least 0.07 m which is larger than 60% of the bed diameter (Grace, 1982). The transition velocity to turbulent fluidization is about 1.1 m/s in our system (Yerushalmi and Cankurt, 1979), much larger than the maximum superficial velocity employed here. It was confirmed visually that our fluidized bed was bubbling for U = 0.037 m/s and slugging for U = 0.095m/s and higher. The slugging pattern was axisymmetric for all size distributions. 6.1 Mean Pressure in the Fluidized Bed The mean pressure, calculated as the average of the instantaneous pressures recorded at a tap 27 or 180 mm above the distributor over a 20-second period, is given in Tables 6.1 and 6.2. The effects of superficial velocity and mean particle diameter on the mean pressure depend on whether the pressure measurement was taken in the 'freely bubbling 104 Chapter 6. Results and Discussion: Pressure Fluctuations 105 Table 6.1: Mean Pressure from Fluidization of Different Size Distributions of F C C . Pres-sure Tap is 27 mm above the Distributor. Bracketted values are the deviations from the mean ror 95% confidence level. Size Distribution Mean Pressure At Different Superficial Gas Velocity (* 103 kPa) U=0.037m/s U=0.095m/s U=0.175m/s U=0.267m/s Original 3.87 (±0 .01) 3.87 (±0 .01) 3.85 (±0 .01 ) 3.81 (±0 .03) Intermediate 3.89 (±0 .01) 3.89 (±0 .01) 3.89 (±0 .02) . 3.91 (±0 .02) Wide 3.89 (±0 .01) 3.88 (±0 .01) 3.87 (±0 .02 ) ' 3.81 (±0 .03) Bimodal 3.87 (±t).01) 3.85 (±0 .02) 3.86 (±0 .02 ) 3.84 (±0 .02) Chapter 6. Results and Discussion: Pressure Fluctuations 106 Table 6.2: Mean Pressure from Fluidization of Different Size Distributions of F C C . Pres-sure Tap is 180 mm above the Distributor. Bracketted values are the deviations from the mean for 95% confidence level. Size Distribution Mean Pressure At Different Superficial Gas Velocity (*10 3kPa) U=0.037m/s U=0.095m/s U=0.175m/s U=0.267m/s Original 2.68 (±0 .01) 2.74 (±0 .02) 2.80 (±0 .04 ) 2.82 (±0 .03) Intermediate 2.71 (±0 .01) 2.75 (±0 .02) 2.84 (±0 .04 ) 2.91 (±0 .03) Wide 2.65 (±0 .01) 2.70 (±0 .01) 2.77 (±0 .04 ) 2.88 (±0 .01) Bimodal 2.70 (±0 .01) 2.73 (±0 .01) 2.79 (±0 .04) 2.88 (±0 .03) Chapter 6. Results and Discussion: Pressure Fluctuations 107 zone' or the 'slugging zone', i.e. the lower or upper portions of the bed, respectively. In the 'freely bubbling zone', a small change in the mean particle size had no dis-cernible effect on the mean pressure. Significant changes in mean pressure in response to the rise in superficial velocity was found for wide distribution and original F C C pow-ders. The three distributions of similar mean particle diameter (intermediate, bimodal and wide) did not show significant difference in mean pressure except perhaps when the superficial velocity was 0.259 m/s. When the bed pressure is measured higher up the column, the mean pressure increases significantly with the superficial velocity as predicted by Svoboda et al. (1984) and Fan et al. (1981) for every F C C size distribution studied. The mean pressure measured when fluidizing the original distribution was significantly higher than that of the wide distri-bution when the superficial velocities were 0.037 and 0.095 m/s. This difference became insignificant when the superficial velocity increased to 0.175 or 0.267 m/s. It seems that by increasing the surface-volume mean particle size, the change in mean pressure in re-sponse to a change in superficial velocity goes up. The particle size distribution again did not seem to affect the mean pressure in a consistent manner. 6.2 Magnitude of Pressure Fluctuations The magnitude of the pressure fluctuation can be represented by the root mean square deviation from the mean pressure. Figures 6.1 and 6.2 show the magitude of pressure fluctuations at different bed levels when fluidizing different F C C distributions. The magnitude of pressure fluctuation increases significantly with superficial velocity. When the bed was merely bubbling, the amplitude and frequency of the pressure waveform were quite irregular as shown in Figure 6.3. The amplitudes of the pressure fluctuations were insensitive to changes in particle size distribution and small changes in mean particle Chapter 6. Results and Discussion: Pressure Fluctuations 108 Cs] C M LEGEND • - ORIGINAL SIZE DISTRIBUTION A - INTERMEDIATE SIZE DISTRIBUTION o - BIMODAL SIZE DISTRIBUTION v = WIDE SIZE DISTRIBUTION to cr 2 ZD •• t—1 ° O L J ° ct: z> cn m U J o ct: ° Cl-lO o o o CM O o o 0.00 0.03 0.06 0.09 0.12 0.15 0.18 0.21 0.24 SUPERFICIAL GAS VELOCITY (m/s) 0.27 0.30 Figure 6.1: Pressure Fluctuations when Fluidizing Different F C C Powders Measured 27 mm above the Distributor. Chapter 6. Results and Discussion: Pressure Fluctuations 109 CO CM O CM OO O CO Q_ cn O ZD • f-r O CJ UJ 0 or CD O CD O O O CM O O O LEGEND ORIGINAL SIZE DISTRIBUTION INTERMEDIATE SIZE DISTRIBUTION BIMODAL SIZE DISTRIBUTION WIDE SIZE DISTRIBUTION 1 1 1 1 1 1 1 1 1 0.00 0.03 0.06 0.09 0.12 0.15 0.18 0.21 0.24 0.27 0.30 SUPERFICIAL GAS VELOCITY (m/s) Figure 6.2: Pressure Fluctuations when Fluidizing Different F C C Powders Measured 180 mm above the Distributor. Error bars are the spread for 95% confidence level Chapter 6. Results and Discussion: Pressure Fluctuations 110 1 1 1 i— • 55Pc OLT (a) Bubbling Regime; Probe at 27 or 180 mm above Distributor. 55 Pa J L . T (b) Slugging regime; Probe at 27 mm above Distributor. 55 Po 1 T 1 1 3 s E (c) Slugging Regime; Probe at 180 mm above Distributor. Figure 6.3: Typical Fluidized Bed Pressure Waveforms for F C C Distributions at Different Bed Levels. Each vertical interval represents a pressure of about 55 Pa. Chapter 6. Results and Discussion: Pressure Fluctuations 111 size. However, the magnitude of the pressure fluctuations did depend on the particle size distribution and mean particle size in the slugging regime. The pressure waveforms also became more regular. The principal differences in size distribution between the wide and the original F C C powders were the higher fines content and larger mean particle diameter of the latter. Decreasing the mean particle diameter significantly lowered the magnitude of fluctuation in a slugging bed when the probe was 27 mm above the distributor. This difference did not persist when the pressure probe was placed at 180 mm above the distributor. Changing the particle size distribution affected the magnitude of the pressure fluctu-ation only when the pressure probe was placed higher up the column. When the pressure tap was very close to the distributor, it seems that only the mean particle size was im-portant in determining the amplitude of the pressure fluctuation. When the probe was placed 0.18 m above the distributor, the particle size distribution became important. At this height, the bubbles had a chance to grow and bring about slugging. A powder that had a narrow particle size distribution produced a significantly higher amplitude of pressure fluctuation than the wide distribution. The magnitude of pressure fluctuations . for the bimodal distribution did not differ significantly from that for either the interme-diate fraction or the the wide distribution. So a powder with a wide and continuous size spectrum provides the least chance of structual damage to the fluidized bed as a result of pressure fluctuations in a slugging bed. 6.3 Frequency of Pressure Fluctuation The frequency of pressure fluctuation does not seem to be influenced by the mean particle size, the particle size distribution, the placement of the pressure probe or the superficial Chapter 6. Results and Discussion: Pressure Fluctuations 112 gas velocity (see Tables 6.3 and 6.4). The frequency was about 1.4 Hz under all the conditions used in this work. There was no definite shift in the frequency of the pressure fluctuation when the pressure probe was moved away from the distributor. 6.4 Summary The mean pressure in the slugging zone increases with increasing superficial velocity. Changing the particle size distribution does not have a definite effect on the mean pres-sure. In the bubbling regime, the magnitude of pressure fluctuation is relative small and is not affected by particle size distribution and small changes in mean particle size. In a slugging bed, the magnitude of pressure fluctuation in the 'slugging zone' is lower for a powder with a wide and continuous particle size distribution than one with a narrow size spectrum. The magnitude of pressure fluctuation increases with superficial velocity regardless of the bed level and particle size distribution. The frequency of the pressure fluctuation did not depend on any of these factors. Chapter 6. Results and Discussion: Pressure Fluctuations 113 Table 6.3: Frequency of Pressure Fluctuation for Fluidization of Different Size Distribu-tions of F C C . Pressure Tap is 27 mm above the Distributor. Size Distribution Frequency of Pressure Fluctuation (Hz) U=0.037m/s U=0.095m/s U=0.175m/s. U=0.267m/s Original 1.47 1.33 1.80 1.58 Intermediate 1.43 1.40 1.43 1.47 Wide 1.27 1.67 1.43 • 1.57 Bimodal 1.80 1.57 1.70 1.47 Chapter 6. Results and Discussion: Pressure Fluctuations 114 Table 6.4: Frequency of Pressure Fluctuation for Fluidization of Different Size Distribu-tions of F C C . Pressure Tap is 180 mm above the Distributor. Size Distribution Frequency of Pressure Fluctuation (Hz) U=0.037m/s U=0,095m/s U=0.175m/s U=0.267m/s Original 1.40 1.33 1.53 1.43 Intermediate 1.40 1.47 1.47 1.40 Wide 1.47 1.40 1.53 1.43 Bimodal 1.50 1.33 1.47 1.50 .Chapter 7 Conclusions and Recommendations A summary on the effects of particle size distribution on fluidization properties is given in Table 7.1. The glass bead and F C C powders have very different fluidization hydro-dynamics from minimum fluidization to vigorous bubbling regimes. For both types of powder, when the mean particle size increases, the Umf, Umb, Ud and fractional bubble free bed expansion increase, but the e^/, emb and ed decrease. Compared to a narrow size cut fraction, a broad size cut decreases emb, increases fractional bubble-free bed ex-pansion and affects Umb- A broad size spectrum FCC powder has higher Ud and lower ed than a narrow fraction. No significant difference in dense phase properties is found between the bimodal and wide F C C distributions. The effect of particle size distribution on the dense phase properties of glass beads is unclear because of the large difference in em/ among the size distributions studied. Increasing the static bed height did not have any significant effects on the minirhum fluidization or the minimum bubbling properties. However, it decreased the superficial dense phase velocity and the dense phase voidage of F C C . With increasing superficial velocity, the mean pressure in the slugging zone of a flu-idized bed increased. The magnitude of pressure fluctuations in the slugging regime increased with increasing superficial velocity and for powders of narrow size cut. Further studies should be conducted to attain deeper understanding of the hydro-dynamics of fluidization involving fine particles. It was found that the physical char-acteristics of powder play a role in the determination of certain fluidization properties. 115 Chapter 7. Conclusions and Recommendations 116 Table 7.1: Summary for the Effect of Particle Size Distribution on the Fluidization Properties of F C C and Glass Beads. FCC Glass Beads Interna. Wide Bimodal Interm. Wide Bimodal H (=Bimodal) L H (=Lnterm) H M L *»/ same same same H M L H L (=BimodaI) L (=Bimodal) H L (=Wide) L (=Interm) H • L M M L H L same H H L (=Bimodal) L (=Wide) umbium, L H same L (=Wide) L (=lnterm) H ud L H (=Bimodal) H (=Wide) H (=Wide) H (=Interm) L H L (=Bimodal) L (=Wide) H M L A ^ r m . at bubbling zone same same same - - -A ^ r m . at slugging zone H L same - -Interm - Intermediate (Narrow Cut) H - High | M - Medium > L - Low same - no statistically significant difference compared with the other two distributions Chapter 7. Conclusions and Recommendations 117 Other powders such as alumina can be used in order to investigate the patterns that these fluidization properties may follow. The experiments here were performed under ambient conditions which were fairly constant throughout the course of investigation. Most industrial fluidized beds are run under different conditions. It would be worthwhile to repeat these experiments at different pressure, temperature and humidity. Only the bubble-free fluidization, bubbling and slugging regimes have been studied in this work. Further experiments should be performed to study the hydrodynamics of fine particle fluidization in the turbulent regime or in a circulating bed. Another fluidization column of different material for the fluidized bed should be built so that the pressure fluctuations can "be studied at different levels of the bed. Nomenclature Archimedes number. Bed diameter, m. Mean particle diameter obtained from standard sieve analysis, m. Mean opening diameter of adjacent sieves, m. Surface-volume mean particle diameter, m. Number-volume mean particle diameter, m. Elasticity coefficient. Mass fraction of powder with particle diameter less than 22/xm. Mass fraction of powder with particle diameter less than 45/xm. Gravitational constant, 9.8 m/sec 2 . Bed Level, m. Bed Height, m. Dense phase bed height,m. Bed height at minimum fluidization, m. Bed Height at minimum bubbling, m. Static bed height, m. Particle Reynolds number. Terminal particle Reynolds number. Superficial gas velocit}', m/s . Superficial dense phase gas velocity, m/s . I 118 | Nomenclature 119 umJ - Minimum fluidization velocity, m/s. Umb - Minimum bubbling velocity, m/s. ut - Particle terminal settlirig velocity, m/s. X - Volume of water to cake 1 kg of powder, m 3 /kg - Weight fraction of powder collected in sieve i. 13 - Exponential proportionality factor. 7 - Exponential proportionality factor. e - Bed voidage. e<f - Dense phase voidage. £mb - Minimum bubbling voidage. Cm] - Minimum fluidization voidage. P- - Gas viscositj', kg/(m*s2). Pg - Gas density, kg/m3 . Pr - Particle density, kg/m3 . P* - Skeletal density, kg/m 3 T - Shear stress. References Abrahamsen, A. R., and D. Geldart, Behaviour of gas-fluidized beds of fine powders part I: homogenous Expansion, Powder Technology, 26 (1980) 35-46. Abrahamsen, A. R., and D. Geldart, Behaviour of gas-fluidized beds of fine powders part II: voidage of dense phase in bubbling beds, Powder Technology, 26 (1980) 47-55. Baeyens, J., and D. Geldart, Predictive calculations of flow parameters in gas flu-idized beds and fluidization behaviour of various powders, La Fluidisation et ses Applications, 263-273, Societe Chimie Industrielle. Toulouse, 1973. Barreto, G. F., J. G. Yates and P. N. Rowe, The effect of pressure on the flow of gas in fluidized beds of fine powders, Chemical Engineering Science, 38 (1983) 1935-1945. Bohle, W., and W. P. M. Swaay, The fluidization of gas adsorption on mass transfer and gas mixing in a fluidized bed, Fluidization - Proceedings of the Second Engineering Foundation Conference, eds. J. F. Davidson and D. L. Keairns, 166-172, Cambridge University Press, Cambridge, 1978. Clift, R., Particle-Particle Interactions in Gas-Particles Systems, Institution of Chemical Engineers Symposium Series, 91 (1986) 27-43. Darton, R. C , R. D. Lanauze, J. F. Davidson and D. Harrison, Zubble growth due to coalescence in fluidized bed, Trans. Inst. Chem. Eng., 55 (1977) 274-280. 120 References 121 Davidson, J. F., and D. Harrison, Fluidised Particles, Cambridge Universitj' Press, 1963. Davies, L., and J. F. Richardson, Gas interchange between bubbles and the contin-uous phase in a fluidised bed, Trans. Inst. Chem. Eng., 44 (1966) T293-T305. Diekman, R., and W. L. Forsythe, Laboratory prediction of flow properties of flu-idized solids, Industrial and Engineering Chemistry, 45 (1953) 1174-1177. Donsi, G., B. Formisani, R. Valentino and G. Volpicelli, The measurement of char-acteristic angles of powders in their prediction in the gas fluidized state, powder Technology, 37 (1984) 39-47. Dry, R. J., M. R. Jugg and T. Shingles, Bubble velocities in fluidized beds of fine dense powders, Powder Technology, 34 (1983) 213-223. Dry, R. J., I. N. Christensen and G. C. Thomas, The effect of naturally-occurring in-terparticle forces on the fluidisation characteristics of fine iron-oxide powders, Chem-ical Engineering Science, 43 (1987) 1033-1038. Ergun, S., Fluid flow through packed column, Chemical Engineering Progress, 48 (1952) 89-94. Ergun, S., Determination of particle density of crushed porous solids, Analytical Chemistry, 23 (1951) 151. Fan, L. T., T. Ho, S. Hiraoka and W. P. Walawender, Pressure fluctuation in a fluidized bed, AIChE Journal, 27 (1981) 388-396. Fan, L. T., T. Ho and YV. P. Walawender, Measurements of the rise velocities of bub-bles, slugs and pressure waves in a gas-solid fluidized bed using pressure fluctuation References 122 signals, AIChE Journal, 29 (1983) 33-39. Foscolo, P. U., L. G. Gibilaro and S. P. Waldram, A unified model for particulate expansion of fluidised beds and flow in fixed porous media, Chemical Engineering Science, 38 (1983) 1251-1260. Foscolo, P. U., and L. G. Gibilaro, A fully predictive criterion for the transition between particulate and aggregate fluidization, Chemical Engineering Science, 39 (1984) 1667. Frantz, J. F., Minimum fluidization velocities and pressure drop in fluidized beds, Chemical Engineering Progress Symposium Series, No. 62, 62 (1966) 21-31. Geldart, D., The effect of particle size and size distribution on the behaviour of gas-fluidised beds, Powder Technology, 6 (1972) 201-205. Geldart, D., Types of gas fluidization, Powder Technology, 7 (1973) 285-292. Geldart, D., and A. R. Abrahamsen, Homogenous fluidization of fine powders using various gases and pressures, Powder Technology 19 (1978) 133-136. Geldart, D., N. Harnby and A. C. Wong, Fluidization of cohesive powders, Powder Technology, 37 (1984) 25-37. Geldart, D., and A. C. Y. Wong, Fluidization of powders showing degrees of cohe-siveness - II: experiments on rate of de-aeration. Chemical Engineering Science 40 (1985) 653-661. Geldart, D. ed., Gas Fluidization Technology, Wiley, New York, 1986. Grace, J. R., Chapter 8 in Handbook of Multiphase System, ed. G. Hetsroni, Hemisphere Press, 1982. References 123 Grace, J. R., Contacting modes and behaviour classification of gas-solid and other two-phase suspension, Canadian Journal of Chemical Engineering, 64 (1986) 353-363. Groot, de, Scaling-up of gas-fluidized bed reactors, Proceedings on International Symposium of Fluidization, ed. A. A. H. Drinkenberg, Netherlands University Press, Eindhoven, 348-361,1967. Jacob, K. V., and A. W. Weimer, High pressure synthesis gas fluidization of fine carbon carbon powders: particle expansion and bubbling, AIChE Journal, 33 (1987) 1698-1706. Jong, J. A. H. de, and J. F. Nomden, Homogenous gas-solid fluidization, Powder Technology, 9 (1974) 91-97. Kang W. K., J. P. Sutherland and G. L. Osberg, Pressure fluctuations in a flu-idized bed with and without screen cylindrical packings, Industrial and Engineering Fundamentals, 6 (1967), No. 4, 499-504. Kehoe, P. W. K., and J. F. Davidson, Continuously slugging fluidised beds, Institu-tion of Chemical Engineering Symposium Series, 33 (1970) 97-116, Butterworth of Australia. Khoe, G. K., T. Ip and J. R_ Grace, The influence of fine particles on the rheology and fluidization behaviour of powders, Proceedings of the 1988 Conference on Measurement and Control of Granular Material, Shenyong, China, 1988. Khoe, G. T., Private communication, 1988. Kmiec, A., Equilibrium of forces in a fluidized bed - experimental verification, Chem-ical Engineering Journal, 23 (1982) 133-136. References 124 Kurtz, R. G., Minimum fluidization velocities: Fluid cracking catalyst and spherical glass beads, Powder Technology, 4 (1970) 156-162. Lirag, R. C., and H. Littman, Statistical stud}' of the pressure fluctuations in a fluidized bed, AIChE Symposium Series, No. 116, 67 (1970) 11-22. Littman, H., and G. A. Homolka, Chemical Engineering Progress Symposium Series, No. 105, 60 (1970) 37. Matheson, G. L., W.A. Herbst and P. H. Holt, Characteristics of fluid-solid system, Industrial and Engineering Chemistry, 41 (1949) 1099-1104. May, W. G., Fluidized-bed reactor studies, Chemical Engineering Progress, 55 (1959) No. 12, 49-56. Massimilla, L., G. Donsi and C. Zucchini, The structure of bubble-free gas fluidized beds of fine fluid cracking catalyst particles, Chemical Engineering Science, 27 (1972) 2005-2015. Massimilla, L., and G. Donsi, Cohesive forces between particles of fluid-bed catalysts, Powder Technology, 15 (1976) 253-260. Mori, S., and C. Y. Wen, Estimation of bubble diameter in gaseous fluidized beds, AIChE Journal, 21 (1975) 109-115. Morse, R. D., and CO. Ballou, The uniformity of fluidization - its measurement and use, Chemical Engineering Progress, 47 (1951) 199-204. Murfitt, P. G., and P. L. Bransbj', Deaeration of powders in loppers, Powder Tech-nology, 27 (1980) 149-162. References 125 Mutsers, S. M. P., and K. Rietema, The effect of interparticle forces on the expansion of a homogenous gas-fluidized bed, Powder Technology, 18 (1977) 239-248. Noodergraaf, I. W., A. van Dijk and C. M. van den Bleek, Fluidization and slugging in large-particle systems,. Powder Technology, 52 (1987) 59-68. Pyle, D. L., and D. Harrison, An experimental investigation of the two- phase theory of fluidization, Chemical Engineering Science, 22 (1967) 1199-1207. Richardson, J. F., and W. N. ZaM, Sedimentation and fluidization: part I, Trans. Instn. Chem. Engrs., 32 (1954) 35-53. Richardson, J. F., Chapter 2 in J. F. Davidson and D. Harrison (Eds.), Fluidiza-tion, Academic Press, London, 1971. Rietema, K., Application of Mechanical Stress Theory to Fluidization, Proceedings of International Symposium on Fluidization, ed. A. A. H- Drinkenberg, 154-175, Netherlands University Press, Eindhoven, 1967. Rietema, K., Powders, what are they, Powder Technology, 37 (1984) 5-23. Rowe, P. N., and C. X. R. Yacono, The bubbling behaviour of fine powders when fluidized, Chemical Engineering Science, 31 (1976) 1179-1192. Rowe, P. N., L. Santoro and J. G. Yates, The division of gas between bubble and interstitial phases in fluidised beds of fine powders, Chemical Engineering Science, 33 (1978) 133-140. Sadasivan, S., D. Bareteau and C. Laguerie, Studies on frequency and magnitude of fluctuations of pressure drop in gas-solid fluidised beds, Powder Technology, 26 (1980) 67-74. References 126 Satija, S., and L. Fan, Characteristics of slugging regime and transition to turbulent regime for fluidized beds of large coarse particles, AIChE Journal, 31 (1985) 1554-1562. , Simone, S., and P. Harriott, Fluidization of fine powders with air in the particulate and the bubbling regions, Powder Technology, 26 (1980) 161-167. Stewart, P. S. B., and J. F. Davidson, Slug flow in fluidised beds, Powder Technology, 1 (1967) 61-80. Stockham, J. D., and E. G. Fochtman, Particle Size Analysis, Ann Arbor Science, Ann Arbor, 1978. Svoboda, K., J. Cermak, M. Hartman, J. Drahos and K. Selucky, Influence of particle size on the pressure fluctuations and slugging in a fluidized bed, AIChE Journal, 30 (1984) 513-517. Toomey, R, D., and H. F. Johnstone, Gaseous fluidization of solid particles, Chemical Engineering Progress, 48 (1952) 220-226. Tung Y., Z. Ziping, Y. Xia, W. Zheng, Y. Yang and M. Kwauk, Assessing fluidizing characteristics of powders, in Fluidization, eds. J. R. Grace and L. W. Shemilt, Engineering Foundation, New York, 1989. Verloop, J., and P. M. Heertjes, Periodic pressure fluctuation in fluidized beds, Chemical Engineering Science, 29 (1974) 1035-1042. Volk, W., C. A. Johnson and H. H. Stotler, Effect of reactor internals on qualitj' of fluidization, Chemical Engineering Progress 58 (1962) 38-41. References 127 Vries, R. J. de, W. P. M. van Swaaij, C. Mantovani and A. Heijkoop, Proceedings of Fifth European / Second International Symposium on Chemical Reaction Engineering, B9-B59, Elsevier Publishing Compan}', Amsterdam, 1972. Wen, C. Y., and Y. H. Yu, Chemical Engineering Progress Symposium Series, 62 (1966) 100. Wen, C. Y., and Y. H. Yu, A generalized method for predicting the minimum flu-idization velocities, AIChE Journal, 12 (1966) 610. Yershalmi, J., and N. T. Cankurt, Further studies of the regimes of fluidization, Powder Technology, 24 (1979) 187-205. Appendix A. Minimum Fluidization Plots 128 Appendix A. Minimum Fluidization Plots 129 in o d -| . 1 1 1 r -0.000 0.002 0.004 0.006 0.008 0.010 0.012 SUPERFICIAL VELOCITY (m/s) Figure A.l: Minimum Fluidization Plots for Original Distribution of FCC. Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots 130 in o in in L J n CJD m L J o CO O LEGEND INCREASING VELOCITY DECREASING VELOCITY o.ooo — i — 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) —i—— 0.010 0.012 Figure A.2: Minimum Fluidization Plots for Original Distribution of FCC. Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 131 CM in o LEGEND o = INCREASING VELOCITY o = DECREASING VELOCITY 1 1 1 1 1 1 1 0.000 0.002 0.004 0.006 0.008 0.010 0.012 SUPERFICIAL VELOCITY (m/s) Figure A.3: Minimum Fluidization Plots for Coarse Distribution of FCC. Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots 132 in o in UJ CD CC Q o UJ CO LEGEND o - INCREASING VELOCITY o = DECREASING VELOCITY 1 1- 1 I 71 0.000 0.002 0.001 0.006 0.008 0.010 SUPERFICIAL VELOCITY (m/s) 0.012 in to' O 0- ° -X to Q_ 1 0 O or Q o £ ~ CD CO LO to UJ Cd 0_ o o "~ UJ CD in a o o • o o • LEGEND INCREASING VELOCITY DECREASING VELOCITY 1 1 1 1 1 — 0.000 0.002 0.004 0.006 0.008 0.010 SUPERFICIAL VELOCITY (m/s) 0.012 Figure A.4: Minimum Fluidization Plots for Coarse Distribution of FCC. Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 133 in o ° - | 1 1 1 1 1 0.000 0.002 0.004 0.006 0.008 0.010 0.012 SUPERFICIAL VELOCITY (m/s) o Figure A.5: Minimum Fluidization Plots for Intermediate Distribution of FCC. Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots 134 in o in in C J to CD in O "? UJ o CD cn O LEGEND INCREASING VELOCITY DECREASING VELOCITY • • • • • 0.000 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) — i — 0.010 o •w in n " o o_ o n ~ o_ in O Ol" CC Q o UJ CM" CC ZD cn in CO _J -UJ or o_ o a — " UJ CO in o" o o o o am o o o • o o- • o • LEGEND D - INCREASING VELOCITY o - DECREASING VELOCITY • 1— 1 r — 1 0.000 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) 0.010 0.012 0.012 Figure A.6: Min imum Fluidization Plots for Intermediate Distribution of F C C . Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 135 o o_ ' o 0 _ C M " O CC Q in C J CC = 5 CD CO 0 C J OT — Q_ O L J f 0 3 A-- O • O O D D D • o n • o • LEGEND o - INCREASING VELOCITY o - DECREASING VELOCITY o o . T -r 0.000 0.002 0.004 0.006 SUPERFICIAL VELOCITY 0.008 0.010 (m/s) Figure A.7: Minimum Fluidization Plots for Fine Distribution of F C C . Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots 136 to o o to UJ CD CC co Q tn '—' o O I I to CO o in d C J to LEGEND o - INCREASING VELOCITY o - DECREASING VELOCITY 1 1 1 T 0.000 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) o.oio o Q_ ° J£ n 0_ 1 0 O rvj or a o or CD cn m cn UJ or o_ o o UJ CO in o d. LEGEND ~ INCREASING VELOCITY o ° - DECREASING VELOCITY 0.000 0.002 0.004 . . 0.006 SUPERFICIAL VELOCITY (m/s) 0.008 0.010 Figure A.8: Minimum Fluidization Plots for Fine Distribution of FCC. Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 137 Figure A.9: Minimum Fluidization Plots for Wide Distribution of FCC. Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots 138 o " o 0.000 0.002 0.004 0.006 0.008 0.010 0.012 SUPERFICIAL VELOCITY (m/s) Figure A. 10: Minimum Fluidization Plots for Wide Distribution of FCC. Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 139 in o in LJ d" CD CC o o L J CD co O CD O . LEGEND INCREASING VELOCITY DECREASING VELOCITY 0.000 —I 0.002 T T 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) — i — 0.010 0.012 n Figure A.ll: Minimum Fluidization Plots for Bimodal Distribution of FCC. Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots 140 in o UJ o CD cr o o > o UJ CQ co o o. LEGEND INCREASING VELOCITY DECREASING VELOCITY 1 1 1 1 1— 0.000 0.002 0.004 0.006 0.008 0.010 SUPERFICIAL VELOCITY (m/s) 0.012 o i 1 r 1 1 1 1 0.000 0.002 0.004 0.006 0.008 0.010 0.012 SUPERFICIAL VELOCITY (m/s) Figure A.12: Minimum Fluidization Plots for Bimodal Distribution of F C C . Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 141 LEGEND o - INCREASING VELOCITY o 1 o - DECREASING VELOCITY UJ CD cr Q o > Q UJ CO Figure A.13: Minimum Fluidization Plots for Coarse Distribution of Glass Beads. Static bed height is 3.3 times bed diameter. Appendix A. Minimum Fluidization Plots U2 in o 1 1 1 1 1 1 1 0.000 0.002 0.004 0.006 0.008 0.010 0.012 SUPERFICIAL VELOCITY (m/s) Figure A. 14: Minimum Fluidization Plots for Intermediate Distribution of Glass Beads Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 143 o O or o o y to ~ or ZD CO o co _:-_©—-a=e-u n • n n • —S" LEGEND INCREASING VELOCITY DECREASING VELOCITY o 1 1 1 1 1 0.000 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) 0.010 0.012 Figure A. 15: Minimum Fluidization Plots for Fine Distribution of Glass Beads. Static bed height is 3.2 times bed diameter. Appendix A. Minimum Fluidization Plots 144 o ID ' O o_ 0_ LJ CC o cn " en LJ ct: o o L J ^ CO ° LEGEND a - INCREASING VELOCITY o - DECREASING VELOCITY o 0.000 —I 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) — i — 0.010 0.012 Figure A.16: Minimum Fluidization Plots for Wide Distribution of Glass Beads. Static bed height is 4.4 times bed diameter. Appendix A. Minimum Fluidization Plots 145 LEGEND INCREASING VELOCITY DECREASING VELOCITY UJ CD tr. O o UJ CD 0.000 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) 0.012 o CO " o o_ °. _£ to" 0_ <=> O IT) ' cr a o UJ ^ > cr ZD CD o CO ^ -UJ rr o_ o o ™" UJ 00 o o d. u • • o Q ° " n — B - e - e — B - O LEGEND INCREASING VELOCITY DECREASING VELOCITY o.ooo — i — 0.002 0.004 0.006 0.008 SUPERFICIAL VELOCITY (m/s) — i — 0.010 0.012 Figure A.17: Minimum Fluidization Plots for Bimodal Distribution of Glass Beads. Static bed height is 4.4 times bed diameter. Appendix B Terminal Velocities 146 Appendix B. Terminal Velocities 147 Table B.l: Terminal Settling Velocity Correlations for Spheres (Grace, 1986). d^ Range Ui Range Equation <3.8 <0.624 Ut~ = K) 2/18 - 3.1234 * 10-4(d;)5 + 1.6415 * 10"6(ci;)8 - 7.278 * 10-lo(d;)n 3.8 to 7.58 0.624 to 1.63 log10c7t" = -1.5466 + 2.9162a; - 1.0432a;2 7.58 to 227 1.63 to 28 log10c7; = -1.64758 + 2.94786a; - 1.09703a;2 + 0.17129a;3 227 to 3350 28 to 91.7 \ogwUt' = 5.1837 - 4.51034a; + 1.687a;2 - 0.189135a;3 dp = dP * \9 * Pg * (PP ~ Pg)/P2}°-u; = ut * \p2gipg{Pp - PgT™ a. = log(d;) Appendix B. Terminal Velocities 148 Table B.2: Terminal Velocities, Reynolds Numbers and indices 'n' for Different Distri-butions of FCC. The indices are predicted using Richardson and Zaki's correlations (See Section 1.3.2) Size d; Ret- n Distributions (m/s) Original 2.20 0.124 0.560 5.19 Coarse 4.02 0.287 2.395 4.09 Intermediate 2.61 0.168 0.898 4.50 Fines 1.71 0.079 0.270 6.45 Wide 2.58 0.164 0.872 4.54 Bimodal 2.64 0.172 0.927 4.46 Appendix B. Terminal Velocities 149 Table B.3: Terminal Velocities, Reynolds Numbers and indices 'n' for Different Distribu-tions of Glass Beads. The indices are predicted using Richardson and Zaki's correlations (see Section 1.3.2). Size Distributions d; ut (m/s) Ret n Coarse 6.59 0.813 9.18 3.58 Intermediate 4.24 0.437 3.18 3.98 Fines 2.18 0.146 0.54 5.24 Wide 4.20 0.430 3.10 3.99 Bimodal 4.26 0.440 3.22 3.97 Appendix C Raw Data from Collapse Tests 150 Appendix C. Raw Data from Collapse Tests 151 Table C.l: Raw Data from Collapse Tests on Original FCC Distribution. Four runs were performed for each set of data. O r i g i n a l D i s t r i b u t i o n of FCC H = 3.3D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 2.77 0.511 0.002 2.37 0.04 4.26 0.548 0.000 4.04 0.02 5.71 0.555 0.003 3.97 0.17 7.20 0.546 0.001 3.49 0.05 8.69 0.540 0.001 3.18 0.02 10.20 0.535 0.002 3.02 0.04 13.46 0.526 0.002 2.70 0.08 17.81 0.518 • 0.000 2.62 0.04 21.55 0.513 0.002 2.67 0.15 24 .97 0.510 0.001 2.89 0.14 28.90 0.508 0.000 2.89 0.15 34.13 0.504 0.002 2.82 0.10 38.91 0.503 0.003 2.81 0.12 H = 4.4D 2.77 0.510 0.000 2.42 0.01 4.24 0.547 0.001 4.01 0.10 5.73 0.550 0.003 . 4.06 0.14 7.23 0.537 0.007 3.45 0.27 8.76 0.533 0.002 3.30 0.08 10.24 0.525 0.001 2.90 0.03 13.65 0.516 0.000 2.66 0.01 17.96 0.508 0.001 2.41 0.10 21.83 0.501 0.001 2.37 0.06 25.15 0.498 0.000 2.43 0.06 29.20 0.496 0.001 2.41 0.09 34.43 0.493 0.001 2.39 0.09 39.07 0.494 0.000 2.36 0.10 Appendix C. Raw Data from Collapse Tests 152 Table C.2: Raw Data from Collapse Tests on Coarse FCC Distribution. Four runs were performed for each set of data. Coarse D i s t r i b u t i o n of FCC H = 3.3D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 7.29 0.487 0.001 5.39 0.01 8.80 0.501 0.000 6.63 0.04 10.33 0.499 0.000 6.21 0.05 13.55 0.500 0.001 5.87 0.27 17.89 0.492 0.001 5.40 0.11 21.33 0.487 0.001 5.70 0.21 25.11 0.486 0.001 ' 5.89 0.01 29.79 0.484 0.000 6.04 0.14 34.37 0.485 0.001 5.85 0.38 39.24 0.485 0.001 5.97 0.14 H = 4.4D 7.31 0.485 0.000 5.20 0.08 8.82 0.500 0.001 6.60 0.03 10.34 0.497 0.002 • 6.11 0.25 13.72 0.481 0.001 5.59 0.29 18.12 0.474 0.001 5.54 0.21 21.42 0.469 0.001 5.60 0.05 25.45 0.468 0.000 5.75 0.10 29.80 0.469 0.000 5.86 0.10 34.77 0.468 0.000 5.79 0.00 39.60 0.468 0.000 5.95 0.06 Appendix C. Raw Data from Collapse Tests 153 Table C.3: Raw Data, from Collapse Tests on Intermediate FCC Distribution. Three runs were performed for each set of data. Intermediate D i s t r i b u t i o n of FCC H = 3.3D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 4 .27 0.531 0.000 3.60 0.00 5.74 0.545 0.000 4.12 0.07 7.25 0.541 0.000 3.86 0.00 8.70 0.539 0.000 3.76 0.01 10.20 0.532 0.000 3.53 0.03 13.48 0.522 0.001 3.39 0.08 17.77 0.514 0.000 3.16 0.00 21.72 0.512 0.000 3.22 0.03 24.97 0.509 0.002 3.31 0.17 28.92 0.507 0.002 3.24 0.05 H = 4.4D 3.56 0.509 0.001 2.52 0.04 4.30 0.527 0.000 3.53 0.01 5.79 0.537 0.002 3.87 0.10 7.25 0.537 0.001 3.77 0.08 8.73 0.532 0.002 3.50 0.08 10.46 0.525 0.002 3.21 0.06 13.75 0.512 0.000 2.78 0.02 18.17 0.507 0.001 2.69 0.18 22.13 0.505 0.001 2.64 0.03 25.83 0.504 0.001 2.82 0.03 29.24 0.502 0.000 2.73 0.02 34.57 0.501 0.001 2.79 0.10 38.92 0.501 0.001 2.75 0.13 Appendix C. Raw Data from Collapse Tests 154 Table C.4: Raw Data from Collapse Tests on Fine FCC Distribution. Four runs were performed for each set of data. , Fine D i s t r i b u t i o n of FCC H = 3.3D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 7.29 0.487 0.001 5.39 0.01 8.80 0.501 0.000 6.63 0.04 10.33 0.499 0.000 6.21 0.05 13.55 0.500 0.001 5.87 0.27 17.89 0.492 0.001 5.40 0.11 21.33 0.487 0.001 5.70 0.21 25.11 0.486 0.001 5.89 0.01 29.79 0.484 0.000 6.04 0.14 34.37 0.485 0.001 5.85 0.38 39.24 0.485 0.001 5.97 0.14 H = 4.4D 7.31 0.485 0.000 5.20 0.08 8.82 0.500 0.001 6.60 0.03 10.34 0.497 0.002 6.11 0.25 13.72 0.481 0.001 5.59 0.29 18.12 0.474 0.001 5.54 0.21 21.42 0.469 0.001 5.60 0.05 25.45 0.468 0.000 5.75 0.10 29.80 0.469 0.000 5.86 0.10 34.77 0.468 0.000 5.79 0.00 39.60 0.468 0.000 5.95 0.06 Appendix C. Raw Data from Collapse Tests 155 Table C.5: Raw Data from Collapse Tests on Wide FCC Distribution. Four runs were performed for each set of data. Wide D i s t r i b u t i o n of FCC H = 3.3D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 3.51 0.497 0.001 3.06 0.02 4.25 0.509 0.000 3.74 0.04 5.35 0.530 0.000 4.96 0.04 5.70 0.536 0.001 5.00 0.27 7.19 0.531 0.003 4.40 0.15 8.66 0.526 0.001 4.13 0.08 10.20 0.518 0.003 3.68 0.08 13.42 0.508 0.001 3.53 0.03 17.73 0.500 0.001 3.56 0.14 21.67 0.495 0.000 3.59 0.12 24.87 0.491 0.001 3.55 0.19 28.83 0.487 0.001 3.36 0.10 33.92 0.488 0.001 3.40 0.18 38.78 0.488 0.001 3.48 0.10 H = 4.4D 3.57 0.494 0.001 3.00 0.01 4.31 0.506 0.000 3.68 0.03 5.42 0.527 0.001 4.95 0.04 5.78 0.534 0.001 4.61 0.12 7.25 0.525 0.001 4.10 0.20 8.77 0.515 0.002 3.64 0.06 10.27 0.511 0.002 3.43 0.04 13.59 0.496 0.003 3.02 0.13 17.94 0.488 0.000 3.01 0.04 21.83 0.481 0.001 2.96 0.03 25.20 0.480 0.002 3.03 0.06 29.16 0.482 0.001 3.17 0.08 34.38 0.478 0.002 3.08 0.14 39.19 0.479 0.002 3.12 0.07 Appendix C. Raw Data from Collapse Tests 156 Table C.6: Raw Data from Collapse Tests on Bimodal FCC Distribution. Four runs were performed for each set of data. Bimodal D i s t r i b u t i o n of FCC H = 3.3D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 3.52 0.495 0.000 2.97 0.02 4.26 0.507 0.001 3.72 0.06 5.36 0.524 0.001 4 .87 0.04 5.71 0.530 0.001 5.26 0.10 7.18 0.535 0.001 4.82 0.07 8.67 0.528 0.003 4.36 0.23 10:16 0.524 0.001 4.00 • 0.11 13.41 0.511 0.001 3.78 0.16 17.74 0.500 0.001 3.66 0.14 21.55 0.494 0.001 3.61 0.08 24.88 0.493 0.001 3.56 0.16 28.79 0.494 0.001 3.49 0.13 34.01 0.492 0.002 3.56 0.13 38.91 0.491 0.001 3.62 0.17 H = 4.4D-3.50 0.492 0.000 2.84 0.07 4.24 0.503 0.000 3.66 0.02 5.35 0.519 0.001 4.76 0.06 5.10 0.526 0.001 5.20 0.11 7.20 0.529 0.002 4.46 0.23 8.75 0.515 0.001 3.91 0.02 10.23 0.507 0.001 3.56 0.08 13.60 0.502 0.001 3.40 0.14 17.95 0.490 0.000 3.13 0.13 21.80 0.485 0.002 3.11 0.06 25.15 0.485 0.000 3.09 0.08 29.13 0.485 0.002 3.26 0.18 34.40 0.482 0.000 3.20 0.05 39.20 0.482 0.001 3.34 0.20 Appendix C. Raw Data from Collapse Tests 157 Table C.7: Raw Data from Collapse Tests on Intermediate Glass Bead Distribution. Four runs were performed for each set of data. Intermediate D i s t r i b u t i o n of Glass Beads H = 4.4*D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 6.87 0.473 0.000 4.75 0.03 7.65 0.480 0.000 5.42 0.03 9.19 0.488 0.001 5.84 0.08 10.74 0.481 0.001 4.90 0.12 14.17 0.471 0.001 4.14 0.17 18.57 0.466 0.001 3.95 0.29 22.54 0.466 0.001 4.93 0.07 26.00 0.466 0.001 5.08 0.32 30.12 0.464 0.001 5.16 0.24 35.56 0.465 0.001 5.03 0.33 40.60 0.464 0.001 5.05 0.22 Appendix C. Raw Data from Collapse Tests 158 Table C.8: Raw Data from Collapse Tests on Wide Glass Bead Distribution. Four runs were performed for each set of data. Wide D i s t r i b u t i o n of G l a s s Beads H = 4.4*D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 6.11 0.453 0.001 4.75 0.07 7.69 0.470 0.000 5.76 0.00 9.23 0.472 0.001 5.28 0.08 10.78 0.462 0.001 4.47 0.03 14.28 0.455 0.002 3.70 0.06 18.64 0.439 0.000 4.81 0.08 22.50 0.434 0.001 4.98 0.11 26.09 0.432 0.000 4.94 0.15 30.07 0.431 0.001 4.96 0.42 35.63 0.431 0.001 4.96 0.06 40.75 0.431 0.001 5.04 0.02 Appendix C. Raw Data from Collapse Tests 159 Table C.9: Raw Data from Collapse Tests on Bimodal Glass Bead Distribution. Four runs were performed for each set of data. Bimodal D i s t r i b u t i o n of Glass Beads H = 4 . 4*D S u p e r f i c i a l Dense Phase S u p e r f i c i a l V e l o c i t y Voidage Dense Phase (mm/s) V e l o c i t y (mm/s) Mean Standard Mean Standard D e v i a t i o n D e v i a t i o n 4.51 0.418 0.001 3.57 0.02 6.11 0.439 0.001 5.05 0.04 7.65 0.459 0.001 6.30 0.04 9.23 0.457 0.002 5.26 0.25 10.81 0.449 0.001 4.43 0.07 14.29 0.431 0.002 4 .30 0.03 18.69 0.408 0.003 3.34- 0.08 22.73 0.401 0.000 3.61 0.04 2'6.18 0.396 0.001 3.56 0.13 30.44 0.398 0.003 3.55 0.05 35.88 0.395 0.001 3.78 0.21 40.88 0.396 0.002 3.67 0.06 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058825/manifest

Comment

Related Items